8.1 Introduction

Since its birth as a science for the objective observation of languages, linguistics has emerged as a comparative study of different tongues. For example, Indo-European linguistics has been dealing with, since the late eighteenth century, the common origin of some of the languages in Europe and Asia before they had access to writing (for a synthesis on Indo-European linguistics, see Reference Klein, Joseph and FritzKlein et al. 2017), and Romance linguistics has been engaging with, largely by means of written content, the coincidences and dissimilarities of languages descended from Latin. The cultural globalization of the twenty-first century is promoting the comparative approach over the study of isolated systems. This cultural rationale is shown in the current interest in a general and typological theory of writing that describes units in all languages or compares them over time. Thus, comparative and historical graphematics is an emerging field of research, as a discipline within the umbrella of historical orthography that compares the graphic units of languages. Comparison is the first methodological imperative of the following new disciplines that study writing.

1. (a) Comparative graphematics, as it has been proposed by Reference Weingarten, Borgwaldt and JoyceWeingarten (2011: 13–14, 35), “deals with the writing systems of the world. It aims to identify the linguistic parameters that underlie cross-linguistic differences between writing systems […] The analysis of at least two languages can provide greater insight into the relevant linguistic structures than can be derived from the analysis of a single language alone.” Comparative graphematics “does not compare writing systems as a whole” because only “detailed linguistic analyses are possible.” Comparisons of writing systems are important tools for the decipherment of yet unknown writing systems, as the typological studies that determine the universal parameters along with writing systems vary and so do various fields of applied linguistics (see also Reference MeletisMeletis 2022, Reference Meletis and DürscheidMeletis and Dürscheid 2022: 119).

2. (b) The status of the comparative concept is also central in linguistic typology (see Linguistic Typology 2016).

3. (c) Grapholinguistics is an emerging interdisciplinary and general area of research that studies all types of writing, not only alphabetic writing systems, and reflects on the role that writing and writing systems play in neighboring disciplines, such as computer science and information technology, communication, typography, psychology and pedagogy (Reference Neef, Sahel and WeingartenNeef et al. 2012, Reference NeefNeef 2015, Reference HaralambousHaralambous 2019, Reference Haralambous2021, Reference MeletisMeletis 2019a: 27, Reference Meletis2020a: 78, 173, Reference Meletis2020b: 250, 2021: 133–34, Reference Meletis and DürscheidMeletis and Dürscheid 2022). See also Chapter 6, on grapholinguistic approaches.

4. (d) Scripturology is a new discipline established by Reference Klinkenberg and PolisKlinkenberg and Polis (2018: 57–59) which concerns the study of different facets of writing but whose scope is wider than in graphematics or graphemics. Scripturology is understood as “a general theory aiming at the establishment of a semiotic typology of writing systems”; “[…] as a search for the universals of writing, as with the language universals.” That said, for Klinkenberg and Polis describing, classifying and ordering the diversity of systems remains an urgent task and it assumes the possibility of comparison (see also Chapter 6, this volume).

This chapter is intended to offer assistance for the linguistic description of writing systems throughout the history of one or several languages. The first decision of the researcher for a comparative analysis of writing systems is to adopt a valid definition of the grapheme, the core unit of written language. This is an essential methodological step, since different schools of linguistics consider different elements to be graphemes. In the first section, we consequently establish a working definition of the grapheme, paying special attention to English and Romance languages. The application of this concept is subsequently exemplified by three different methods of diachronic and comparative description of writing systems: Romance scriptologie, cultural history of European orthographies and comparative graphematics of punctuation. In the second section of this chapter, we discuss biscriptality, which studies the use of two or more writing systems for representing the same language – a phenomenon that is not rare in the history of languages from different families, and that is related to different aspects of society and language users. With examples mainly from Slavic languages, biscriptality is shown to be present on several levels of written language. Biscriptality situations are characterized by dichotomies such as synchronic vs. diachronic biscriptality, monocentric vs. pluricentric biscriptality, and societal vs. individual biscriptality.

8.2 Comparative Graphematics and the Units of Writing Systems

8.3 Biscriptality

Comparative historical perspectives are always related to different aspects of society and language users. In Section 8.2, we have described different methodologies for diachronic comparison of writing systems of different languages. In the present section, we also deal with the analysis of different graphic sets of signs used for one and the same language.

8.4 Conclusion

Comparison is the first methodological imperative of the new disciplines that study writing: comparisons of writing systems are important tools for the decipherment of yet unknown writing systems, for typological studies that propose to determine universal parameters for the variation between and within diverse writing systems, and for various studies within applied linguistics. Thus, the comparability of orthographies of different languages and societies across time constitutes an emerging field of investigation within international linguistics. An essential methodological step in this field is the definition of the grapheme as the core unit of written language. The grapheme definition determines different methodologies and approaches of comparison, such as, for example, Romance scriptologie, the common cultural history of European orthographies in the period of early modern Europe, or the comparative graphematics of punctuation and orthotypography. If we include punctuation signs in the definition of the grapheme, then comparative graphematics and the comparative cultural history of orthographies may fill in some gaps in the multidisciplinary research on writing systems. Similarly, the renewed scholarly interest in the use of two or more writing systems for one and the same language underscores the fact that biscriptality – with differing definitions and in a multitude of perspectives – constitutes an important feature of many language situations throughout the history of writing.

9.1 Introduction

We may begin with the following quotations, which are useful springboards for the issues discussed in this chapter:

The passage quoted from the first century AD Roman author Quintilian’s treatise on public speaking argues for what many surely regard as a self-evident truth about the relationship between spelling and pronunciation: the former should accurately reflect the latter. He is not dogmatic about it, however, as his parenthetical insert nisi quod consuetudo optinuerit ‘except where usage prevails’ makes clear: he allows custom to override this principle even if causing deviations from it.Footnote ¹ If the pseudo-spelling ghoti for fish in the second quote above (later made famous by George Bernard Shaw, see also Chapter 31) in fact had the weight of customary usage behind it, it might occasion some consternation on the part of those learning how to spell, but once it became fixed in the memory and used countless times, few if any people in the English-speaking world would probably bat an eye over it. The question of when customary usage should be given the green light to buck otherwise systematically applied orthographic principles is old and vexed, and will likely never find true resolution because of an essential fact about writing systems (by which is meant not just the written symbols per se but also the principles of their use and combination): contrary to Quintilian’s stated goal, systems of spelling have never in practice behaved like mathematical functions or computer algorithms where a single output is mechanically, unconditionally and exceptionlessly generated from a particular input – or, to put it less abstractly, where a particular unit of language is exceptionlessly and uniquely signaled by a particular written symbol. This chapter is a brief exploration of some of the ways in which writing ‘systems’ are and are not, in fact, systematic, and some of the ways in which idiosyncrasies, irregularities, inconsistencies and ambiguities can crop up, especially in the mapping of speech onto writing.

By way of outlining what is to come, after some further preliminaries, the first subject treated is graphemes in a given writing system as entities separate from issues of their orthographic use, with focus on those that form an internally coherent set of symbols as regards their construction, visual appearance or organization. After this, the relationships between graphemes or orthographic systems and speech sounds (both phonemes and allophones) are treated, with abundant exemplification of various degrees and levels of systematicity and idiosyncrasy. The discussion then advances to the level of whole words and their spelling, before a brief conclusion. The topics naturally overlap to some extent and should not be taken as too strictly imagined; some of the material under one heading is relevant to other headings as well. The examples, collected from a wide variety of languages and scripts, are selective and specific and, even when going into detail, are not intended to provide comprehensive descriptions of any of the relevant orthographic systems, which are treated elsewhere in this volume and in other works.Footnote ² The reader is assumed to have some knowledge of basic terms pertaining to writing systems (such as grapheme, pictogram, logogram and syllabary, as discussed in Chapter 2 and Chapter 3, this volume).

9.2 Sources of Inconsistency and Idiosyncrasy

At the most basic physical level, audible speech is an acoustic signal consisting of a continuous series of pressure waves, while writing consists of static and discrete visible symbols: no writing system has any direct physical connection to speech. The discovery that the acoustic signal – and/or the articulatory gestures that produce it – is composed of recurring units (be they words, morphemes, syllables or individual speech sounds) and that these units can be indexed in a completely different medium by a set of arbitrarily devised visual symbols was a brilliant act of analytic genius, no less so for having been achieved multiple times in human history.Footnote ³ One might indeed think that the creation of such symbols to represent linguistic units would result in a neat, one-to-one system, where each visual symbol always corresponded to one and the same linguistic unit. For various reasons, though, this has been rarely accomplished. For example, many writing systems started out as pictographic, but one cannot draw a picture of everything; to represent abstract concepts, metaphoric and metonymic usage of pictograms had to be introduced, resulting in some symbols becoming polysemous. Thus, in Sumerian cuneiform, for example, the symbol for the word for ‘heaven’ (an), in origin a picture of a star, came to be used logographically also for the word for ‘god’ (diĝir) by metonymic extension (), whereas ‘star’ or ‘constellation’ (mul ) was represented by a different pictogram consisting of three stars (i.e. ). This strategy, too, only takes one so far, and it is in any event not possible to devise a logogram for every word in a given language; the set is unbounded, since humans have an unlimited ability to create novel words.

For this reason, all systems that started out as pictographic or logographic evolved to include a phonetic component so that words could be represented not only according to what they meant but also according to what they sounded like. This means that many signs came to be used both as pictograms/logograms and as phonograms, further departing from any one-symbol, one-meaning principle. This was the course of development seen in, for example, Mesopotamian cuneiform, Egyptian hieroglyphics, Chinese writing and Mayan hieroglyphics. In the case of purely phonetically based writing systems – syllabaries and alphabets – we might expect that it would be easier to achieve thoroughgoing one-to-one correspondence between written symbol and speech sound or sound sequence. In practice, though, this is not the case here, either. Although total regularity can in theory be achieved for a given variety of speech at a particular moment, pronunciation naturally varies at any given time within a speech community, meaning the spelling system might be slightly ‘off’ for some speakers right from the time of the introduction of the writing system. Furthermore, the ineluctable effects of sound change slowly increase the distance between the static written norm and the spoken form in any case. Modern English spelling is an immediate and familiar example of this: many words, such as knight, make and listen, contain so-called ‘silent letters’ that were actually pronounced at the time their spelling became fixed hundreds of years ago, but have since disappeared from speech. Even if spelling at the level of the word could be perfectly regular and systematic, difficulties of implementation arise at higher levels: if, for example, the writing system signals word divisions, should a particular short idiomatic phrase or compound be written as a single word or not (as with smart phone vs. smartphone)?Footnote ⁴

The ever-widening gap between speech and writing over time has of course resulted in reform efforts to bring orthographies in line with pronunciation after the latter has changed, but these are not always consistently implemented. In theory, the gap could be repaired every few generations, but the problem never really goes away; the histories of countless spelling systems, including notably English, are littered with the bones of successive layers of reformation and attempts thereat.Footnote ⁵ But conservative cultural forces – not to mention (especially in modern times) economics and practical realities – can prevent reform from being implemented in whole or in part. Noah Webster’s long-term campaign to regularize American English spelling, which had a strongly moralistic underpinning relating to views of language and national character (see Reference BynackBynack 1984: 104–6), succeeded in a few straightforward cases where he advocated simpler variants already in existence (-or for -our as in color, -er for -re as in center, etc.) but otherwise failed, as people balked at changes that appeared too drastic, such as masheen for machine.Footnote ⁶ The set of orthographic reforms instituted in Germany in 1996 – nearly a century after the previous set of reforms in 1901 – to make spelling more ‘logical’ evoked a hue and cry from educators and prominent intellectuals; implementation wound up proceeding differently in different places and not to the same extent everywhere, and now, nearly a quarter-century later, unresolved issues with it continue to linger.Footnote ⁷

It should be recalled that the purpose of writing systems is first and foremost communicative: they are a means for transmitting information. This can be readily effected without representing all the potentially relevant linguistic units of a language and/or representing them consistently. Native speakers can recover much from context that is not directly expressed in written form. This, plus the factors outlined above that can lead to irregularities, leave little room for surprise that writing tends to be imperfectly systematic in the mapping of spoken language.

9.3 Systematicity at the Level of the Grapheme

When all the written symbols of a given writing system are considered together, there are a couple of levels at which they can exhibit systematicity. The first and most visually obvious one is the stylistic or aesthetic level. Many writing systems have undergone the imposition of aesthetic principles that lend most or all of the written symbols a unified look. This has generally involved either the establishment of a small set of constitutive graphic elements out of which most or all of the symbols were built (or were retrofitted to be built out of), or the introduction of a unified styling in the absence of separate elements. The proto-cuneiform used at the end of the fourth millennium BC in Sumer that marked the beginnings of a writing system for Sumerian is a hodgepodge of incised images of varying sizes composed of curved and straight strokes. As a result of changes in writing technology in the third millennium BC, whereby reed styluses cut at an angle came to be used for impressing the signs, the wedge-shaped strokes that give cuneiform its ‘classic’ look became (by default) the norm, and cuneiform symbols came largely to be made out of only three elements: horizontal wedges, vertical wedges and wedges set at roughly a 45-degree angle. Symbols also came to take up more or less the same amount of vertical space in a line of text (though their width varied considerably).

Similarly, the earliest preserved Chinese characters, most of them incised on flat bones or shells such as scapulae and tortoise plastra in the second millennium BC, had some general aesthetic commonalities but differed widely in size; these too evolved into a system comprising a fixed number of standardized constitutive elements (eight distinct basic stroke-types plus numerous additional combining and compound strokes) with every character in a line of text filling the same amount of space (an imaginary square box) regardless of the number of strokes it contained. The shapes of the strokes themselves also underwent stylistic unification under the influence of brush writing, which occasioned variations in stroke thickness depending on the direction of the stroke (horizontal strokes are normally thinner than vertical and certain angled strokes) as well as, for example, the leftward ‘hook’ at the bottom end of vertical strokes.Footnote ⁸ The evolution of the look of the Latin alphabet is comparable: the earliest inscriptions in Latin, such as the seventh century BC Forum Inscription from Rome, look crude and jumbled by comparison with the aesthetic that was to develop over a half-millennium later, where letters on official and monumental inscriptions were evenly spaced and of a uniform height, and the strokes of each letter varied in thickness according to their orientation and ended flared out in serifs. This look, which survives intact today in traditional typography, was also due to influence from brush writing.

Less frequently, the graphemes of a writing system can explicitly and systematically encode information about specific features of the represented linguistic unit, in being constructed out of a common set of parts with linguistic meaning. More than nine-tenths of Chinese characters are combinations of two simpler characters where one is used as an indicator of the semantic domain and the other provides phonetic information, and there is a standardized fixed set of the semantic indicators (so-called ‘radicals’). A few signs in Mesopotamian cuneiform were similarly compounded out of simpler signs as a way of combining semantic with phonetic information. The vast majority of the world’s writing systems, however, do not evince graphemes whose component parts – if distinct component parts they even have – convey information on their own and/or form a group with parallel parts of other graphemes.Footnote ⁹ Not infrequently, characters in a writing system are formed by modifying other characters that have a related phonetic value. In Cyrillic, which is fundamentally a modified and expanded version of the Greek alphabet, the letter Б : /b/ is a modification of В : /v/, which was an adoption of the Greek letter beta after the latter’s ancient pronunciation as /b/ had evolved to /v/. In Brāhmī, several of the characters for aspirated stops (or, more precisely, aspirated stops plus the inherent vowel a) appear to be related to those representing the corresponding unaspirated stops, in that the former were derived from the latter by the addition of a stroke or the latter were derived from the former by the subtraction of a stroke (or a mixture of both). This can still be seen today in some such pairs in Nāgarī, the most widely used descendant of original Brāhmī, such as unaspirated ṭa beside aspirated ṭha and unaspirated pa beside aspirated pha . But only in one writing system are the shapes of the characters explicitly designed to pictorialize oral gestures involved in the articulation of the respective sounds; that system is Korean Han’gŭl. There is some dispute about which models were used by its promulgator, King Sejong (and/or his scholars), traditionally credited with inventing it in the fifteenth century, but the resulting product is clear: in the case of the velars and dentals, for example, the letters depict the position of the tongue viewed laterally as in a modern cut-away depiction of the vocal tract with the lips to the left, so that the velar g ㄱ shows the back part of the tongue raised toward the roof of the mouth, while symbols for the dentals n ㄴ and d ㄷ show the tip of the tongue raised toward the alveolar ridge. Adding a horizontal stroke inside of g and d produces symbols for the aspirated stops k ㅋ and t ㅌ.

Systematicity at the level of the grapheme can also sometimes be seen in the ordering of the symbols constituting a writing system, ordering which is sometimes language-specific. There is no recoverable rationale to either of the two attested orders of letters in the Semitic alphabet,Footnote ¹⁰ but the ordering of at least one major family of scripts, Brāhmī, has been systematically informed by linguistic principles. The character order here is reflective of the highly developed Indian grammatical tradition described in Section 9.5, and proceeds according to phonetic principles, both in general terms – vowel graphemes and other syllabics are grouped together as a series preceding the nonsyllabic sounds – and at numerous levels of detail. The vowel graphemes/syllabics are grouped into short–long pairs, themselves ordered relative to one another according to natural classes. Thus the three characters corresponding to vowels at the corners of the vowel triangle start things off, beginning with mid or low (a : [ǝ] ā) followed by high front and high back (i ī u ū ); next come the symbols for the two pairs of syllabic sonorants (ṛ ṝ ḷ ḹ), rounded off by the graphemes corresponding to the mid vowels and diphthongs again going from front to back (transliterated nowadays as e ai o au but in earlier literature ē āi ō āu, which is actually a better reflection of the phonetic reality). With this last set, matters are additionally interesting because of two other facts. First is a further feature of all these letters: each long member of a short–long pair is the same character as the short member plus an added stroke (thus, e.g., a, ā ). For e/ai and o/au, this may seem puzzling, but e and o are historically monophthongizations of the (short) diphthongs *ai and *au, and in historical times still alternate with them (or their equivalents ay and av with off-glides) in certain contexts; additionally, e o ai au are all metrically long.Footnote ¹¹ Similar systematicity is evinced by the arrangement of the obstruents, which are grouped overall according to manner of articulation (stops, glides, liquids, fricatives); the stops furthermore are divided into groups ordered according to place of articulation from the back of the mouth forward (velar, palatal, retroflex, dental, labial), and within each of these groups they are further arranged according to manner of articulation (plain voiceless first, followed by voiceless aspirated, voiced, voiced aspirated and nasal).

As Buddhism and, with it, the Brāhmī writing system spread eastward, the system’s linguistically informed ordering proved influential on the arrangement of characters in some unrelated writing systems, particularly the two Japanese kana syllabaries and Han’gŭl. The ordering of the Han’gŭl consonant graphemes is similar to that of Brāhmī (in the original 1446 version of the former, k/g kh ng t/d th n p/b ph m j ch s h ʔ l/r z),Footnote ¹² though the vowel symbols come after the consonant ones. The kana syllabaries of Japanese follow the Brāhmī order more faithfully, especially when one takes sound changes into account. These syllabaries begin with the vowel symbols in the Brāhmī order (a i u e o) followed by the k-, s-, t-, n-, h-, m-, y-, r- and w-series;Footnote ¹³ representation of voiced g, z, etc. is achieved by the addition of two small strokes (thus, for example, く ku and ぐ gu, と to and ど do). The position of the s-series, curious at first glance, is due to the fact that it corresponds to the Brāhmī palatal series: the phonemes corresponding to s and z are allophonically palatalized to the sound reflected by sh and j before i, and j in Brāhmī belongs to the palatal series. As for h, it is historically the weakened descendant of earlier p, which is why adding the diacritic for voicing to members of the h-series produces a b-series; thus, for example, は ha (< pa) and ば ba. Since p remained in some positions, a special diacritic was devised to represent it that looks like a degree sign; thus, for example, ぱ pa.

An unusual case of alphabetical order (going hand in glove with a perhaps unique alphabetic genesis) is represented by the Thaana alphabet used since the seventeenth century to write Dhivehi (Maldivian), which consists of letters that are historically the Arabic numerals from one to nine, followed by the corresponding Indian numerals that had been used in the predecessor Brāhmī-derived script, with vowels indicated by letters derived from the vowel diacritics of Arabic (Reference Gippert, Chen and SladeGippert 2013: 96–98). Previously, Dhivehi had been written in a Brāhmī-derived script that ran left to right, which made it cumbersome to mix in Arabic words (written right to left); Thaana solved that problem since it is written just right to left. However, no rationale has yet been uncovered for the assignation of the specific phonetic values to the numerical symbols, or why numbers were chosen in the first place.

9.4 Phonemic Indication

It is often assumed that phonetically based writing systems typically cover all the sounds of a language at the phonemic level.Footnote ¹⁴ This assumption is largely true of the more prominent Western systems, but, as also briefly discussed in Chapter 3, cases where some phonemic distinctions are not represented are far from rare. In writing systems where each written symbol represents a phoneme (i.e. in alphabets) or a syllable (i.e. in syllabaries), decisions have to be made as to how many phonemes or phonemic sequences are to be represented. Pure abjads represent only consonantal, not vocalic, phonemes, and do not, for example, graphically encode differences between single and geminate (double) consonants. Such, historically at least, are the Semitic alphabets like Phoenician, Old Hebrew and Nabataean. Over time, of course, such systems can expand to encode further information: certain consonantal letters can take on the additional role of indicating long vowels (such as the use of aleph, normally representing a glottal stop, to represent ā also; these are the so-called matres lectionis). Syllabaries, although they represent vowels by default, do not necessarily indicate differences in vowel length or the doubling of consonants, such as Linear B, described directly below. Semitic alphabets historically have not had dedicated letters for the short vowel phonemes; diacritics exist for representing short vowels but these are typically not employed outside of sacred texts and some children’s literature.

An extreme case of phonemic underrepresentation is the Linear B syllabary used for writing Mycenaean Greek in the second millennium BC (see, e.g., Reference HookerHooker 1980; see also Chapter 20, this volume). Linear B contained only symbols for V and CV syllables, and the consonant inventory encoded in the script fell considerably short of the number of phonemic distinctions in the language.Footnote ¹⁵ Greek had a three-way distinction between voiceless, voiced and voiceless aspirated stops (e.g. /p/, /b/, /p^h/), but the script only reflected the first two of those and only in the case of the dentals (/t/ and /d/); for labials, palatals and labiovelars, a single set of CV symbols had to be used for all three manners of articulation.Footnote ¹⁶ Furthermore, the script did not distinguish between /r/ and /l/. Although theoretically long vowels could have been indicated with spellings of the CV-V type, the Linear B scribes appear to have had no interest in representing vowel length, and only devised workarounds for syllable-initial consonant clusters, which were indicated with a dummy copy vowel (e.g. ko-no-so for Knossos). As this example shows, gemination of consonants was not indicated either, nor were consonants in syllable codas.

A not dissimilar case, at least in its end result if not in its origins, is provided by the limitations of the Pahlavi alphabet (Reference MacKenzieMackenzie 1971), a modified form of the Aramaic alphabet used to write Middle Persian and some other Iranian languages. Over time, numerous letters merged in form, resulting in only a dozen distinct letters representing twice as many phonemes. The multiple ambiguities that arise from this can hamper scholarly interpretations of Pahlavi texts, but appear not to have caused Pahlavi speakers themselves any difficulty. Indeed, it can almost be stated as a rule that orthographies try to maximize economy; if a particular phonemic distinction does not have to be indicated, it often will not be. For example, Spanish and Italian indicate the position of the word stress (with an accent mark) only if the accent does not fall on the syllable that the productive stress-positioning rules would lead one to predict; thus in Spanish Alcázar (which violates the rule that words ending in -r are normally end-stressed) and Italian àncora, città (which violate the rule that words are normally stressed on the penult) the stress is written in, but not in, for example, Spanish doctor (stressed on the final syllable) or Italian fiore (stressed on the penultimate syllable). This already provides more information than some other orthographies, such as Russian, where stress is very unpredictable but not represented in writing at all. For example, the word мука means ‘flоur’ if stressed on the first syllable but ‘torture’ if stressed on the second, with no distinction in writing. English is not so different; compare project (noun, first syllable stressed) versus project (verb, second syllable stressed) and many other such pairs where the stress difference is not indicated in spelling.

Sometimes the desire to maximize economy can lead to ambiguities and an actual increase in complexity through the rise of so-called ‘positional rules’. Italian has two phonemes, /k/ and /tʃ/, that can each be represented in two partially overlapping ways depending on context: c before a back vowel is always /k/ and before a front vowel, /tʃ/; /k/ before a front vowel is represented by ch, while /tʃ/ before a back vowel is represented by ci, a sequence of letters that in other contexts represents the syllable /tʃi/. Thus cosa ‘thing’ /kɔsa/ ~ cento ‘hundred’ /tʃɛnto/ ~ che ‘that’ /ke/ ~ ciao ‘bye’ /tʃao/ (not *[tʃiao]) but cinque ‘five’ /tʃiŋkwe/. This system arose for a couple of reasons. First, historically the phoneme represented by c had two different pronunciations depending on the context: it sounded like /k/ before the back vowels /ɑ/, /o/ and /u/, but more like /tʃ/ before front vowels (/e/ and /i/), and subsequent changes resulted in these originally predictable variants becoming separate phonemes. Second, this change in the phonemic inventory was not matched by the introduction of any new written symbols (or resuscitation/recycling of ones in disuse, like k), probably because of a desire to preserve c in all its historically justified contexts. This resulted in the letters h and i being marshaled into service with new roles as indicators of the quality of a preceding c. The desire to maximize economy, which would have been most easily satisfied by adopting a one-grapheme-equals-one-phoneme approach, has in this case conflicted with the desire to keep things looking basically as they always did. This desire itself was not consistent; for example, z was introduced as a replacement of Latinate ti after the latter’s pronunciation had changed to [ts] (for more see Reference DemartiniDemartini 2011). Spanish has a similar situation with respect to the voiced counterpart of /k/, namely /g/. This is represented orthographically by g, but g is pronounced /x/ before e or i. To indicate the sequence /ge/ or /gi/, Spanish orthography uses gu instead, with u acting as a ‘buffer letter’ like the h in Italian ch. However, unlike h in Italian, which is always silent, u in Spanish is not, and gu in other contexts represents /gu/ or /gw/ (often weakened to just /w/). Thus to indicate the sequence /gui/ or /gue/ in pronunciation, Spanish orthography makes use of an additional feature, namely, it adds a diaeresis over the u, as in averigüé ‘I found out’ [-guˈe] from the verb averiguar ‘to find out’ [-guˈaɾ].

An especially interesting case of the ambiguities and idiosyncrasies that can arise through the use of positional rules – however systematic the rules themselves are – comes from Classical Old Irish orthography (Reference ThurneysenThurneysen 1946). Here, letters were pronounced differently depending on their position in a word, resulting in some sounds being unambiguously represented and others not. A c, for example, represented /k/ word-initially but /g/ after a vowel; thus cath /kaθ/ ‘battle’, éc /eːg/ ‘death’. To represent a voiceless stop after a vowel, the letter was written double: macc /mak/ ‘son’, accusative plural maccu /maku/. Certain proclitics (unstressed particles that form a phonological unit with a following stressed word) induced a morphophonological set of mutations to word-initial consonants, but because they were written together with their hosts without a break, this sometimes resulted in changes to the spelling of those initial consonants and sometimes not. One of the mutations resulted in the voicing of a word-initial /k/ to /g/. The genitive plural of the definite article, inna or na, is an example of a proclitic that induced this mutation; ‘of the battles’ was pronounced /(i)nə gaθə/ but written (in)nacathae. The internal -c- unambiguously represents /g/. However, ambiguity sets in when the mutation was induced by a full lexeme, since full lexemes were written separately. Such a word was secht /ʃɛχt/ ‘seven’; ‘seven battles’ was pronounced /ʃɛχt gaθə/ but written secht cathae.Footnote ¹⁷ In certain other cases, however, the change in pronunciation was indicated. If the initial consonant was a voiced stop, such as /g/, a word like secht caused the addition of a homorganic nasal, which was represented orthographically: glais ‘locks’ but secht nglais ‘six locks’. Thus in the same context, words beginning with voiceless stops were written morphematically and the reader had to mentally supply the sound changes, while words beginning with voiced stops were written phonetically. In other contexts, though, the situation could be reversed. The numeral cóic ‘five’ caused a following word-initial stop to become a fricative; thus /k/ became /χ/ and /g/ became /ɣ/. While /χ/ was unambiguously represented as ch, /ɣ/ was written g, just like /g/. Thus ‘five battles’ was written cóic chathae while ‘five locks’ was written cóic glais /ɣlaʃ´/. Here it is the words beginning with voiced stops, not voiceless stops, where the reader has to mentally supply the changes.

Many alphabetic and syllabic writing systems use single and double letters or characters in order to signal contrastive (phonemic) length of vowels and/or consonants. In the cuneiform syllabary described above that was used for Hittite, not all possible consonant-vowel-consonant sequences had a dedicated sign in the script. Those that did not thus had to be represented with a CV sign followed by a VC sign, each encoding the same vowel sound, for example, pa-ak for the syllable /pak/. Since the doubling of the vowel is orthographic only, necessitated by the limitations of the script (p-ak and pa-k were not possible because there were no true consonant letters), in order to indicate a phonetically long vowel, an extra V-sign had to be inserted: pa-a-ak. But a double consonant was always significant (ap-pa) and represented gemination,Footnote ¹⁸ since the orthography easily allowed a contrast between that and a-pa with a single intervocalic consonant (see further Reference MelchertMelchert 1994).Footnote ¹⁹

In contrast with modern orthographic systems like those of Italian and Finnish, where geminate consonants are regularly indicated as such, the Hittites were inconsistent, but even Italian and Finnish are not perfect in this regard. In certain syntagms in both these languages where word-initial gemination occurs, the gemination is not signaled orthographically. The case of raddoppiamento sintattico in Italian is one example: certain vowel-final monosyllables trigger gemination of the initial consonant of the following word, but without this being reflected in the spelling; for example, va bene ‘okay’ is pronounced /va bːɛne/ as though va bbene. However, such phrases are spelled with doubling if they have been univerbated, such as frattanto ‘meanwhile’ < fra ‘between’ + tanto ‘as much, still’. A similar situation arises in Finnish, though with a bit less regularity as regards univerbations. Nouns and adjectives in -e like terve ‘well, healthy’ ended historically in a consonant that assimilated to following consonants in a variety of contexts, resulting in a geminate consonant in speech. As in Italian, the initial geminate is not indicated when the words are separate; but, somewhat more complicated than the Italian situation, within single words, gemination is not always indicated either. Case-endings undergoing the change regularly reflect the gemination, as in the partitive singular (terve-ttä < terve + -tä ). However, in a compound like tervetuloa ‘welcome’ (terve + tuloa ‘coming’), pronounced as though tervettuloa, gemination of the initial t of tuloa is not indicated, and whether a word induces gemination is not always predictable—in many cases it is lexically determined. Here Finnish evinces a genuine irregularity in its otherwise very regular spelling system.Footnote ²⁰

Doubling of a letter is not the only way to indicate gemination or length of the sound indicated by that letter. In Dutch, vowel length can be indicated by whether the following consonant is written single or double: single indicates the preceding vowel is long (e.g. maken with /aː/); double, short (e.g. Bakker with short /a/). This system, however, is only used if the consonant is intervocalic; otherwise, the vowel letter is written single or double depending on the length (e.g. short vowel in van, best, word ; long vowel in aan, twee, woorden). The practice of using the following consonant letter to signal length of a preceding vowel is found in the orthographies of other West Germanic languages, including English (whence the contrast between hoping and hopping, for example).Footnote ²¹ Long vowels can also be indicated by the addition of an extra ‘silent’ letter, such as an h (German Ahl ‘awl’) or an e (Low German placename Coesfeld /koːsfɛlt/). The ‘silent e ’ in English that is synchronically interpreted as an orthographic indicator of vowel ‘length’ (as in made vs. mad ) is not isolated either; a comparable usage is found in the Mayan hieroglyphic script (see Reference CoeCoe 1992). In this logosyllabic script, only syllables of the structure V or CV are represented by dedicated signs. There being no unambiguous way to represent word-final consonants, a dummy vowel was used that copied the vowel of the preceding syllable: thus a syllable like pan was written pa-na. To indicate that the vowel in the preceding syllable was long, the dummy vowel i was used instead, so pa-ni indicated /paːn/. It may seem peculiar that these systems and others like them do not make use of unambiguous symbols like diacritics for indicating length, but rather use workarounds based on the number of graphemes or a particular combination of graphemes or the equivalent. Again, it is often not important for the writing cultures in question that all these details be represented at all; and diacritics in general are something of a latecomer historically, though there are a few examples from the ancient world. Oscan, a relative of Latin in ancient Italy, added a modified I and U to its alphabetic inventory around 300 BC that consisted of the old I and U with an added stroke at the top, used to indicate different vowel qualities. Latin inscriptions and manuscripts sometimes made use of the so-called apex to indicate vowel length, a mark that looks similar to a modern acute accent. In both of these traditions, use of these diacritics was optional.Footnote ²²

Besides not always representing the full phonemic inventory or all phonemic distinctions of a language, written cultures have sometimes seen fit to ignore entire morphemes. An extreme example comes from ancient Sumer, where the earliest layer of true writing contained logograms for a variety of full lexical concepts (the root meanings of nouns, verbs and adjectives) but had not yet developed syllabograms to represent inflectional morphology.Footnote ²³ A less extreme example, from the modern world, comes from Farsi (Reference ThackstonThackston 1993). In this language’s so-called ezâfe construction, a clitic particle -e or (after vowels) -ye is inserted between the elements of certain phrases, such as noun phrases (e.g. mard-e xub ‘a good man’, hanâ-ye xub ‘good weather’). Because the language is written in the Perso-Arabic script, which is an abjad and does not indicate short vowels, this grammatical element is completely absent from writing in phrases like the first one above (written مرد خب mrd xb), but it is reflected in phrases like the second one since -ye begins with a consonant (حناط خب hn’y xb, with aleph ا ’ used for â ).

But is phonetic regularity even a desired feature of writing systems? Specialists in language, who are steeped in the International Phonetic Alphabet, might put a premium on writing systems that have a one-to-one relationship between written symbol and sound, but particular writing cultures often put their priorities elsewhere. Reference SebbaSebba (2007: 81–82) notes the interesting cases of Romanian and Moldovan, the name for Romanian as spoken in the Republic of Moldova and, historically, in much of Moldavia (which stretched over what is now part of eastern Romania, Moldova, and parts of Ukraine and Russia). Until relatively recently, the Cyrillic alphabet was used to write Romanian/Moldovan. In early nineteenth-century Romania, the Latin alphabet was introduced on the grounds that this was appropriate given the language’s descent from Latin; Cyrillic was banned in 1863. In Moldavia, Cyrillic continued to be used until the 1920s, when the switch to the Latin alphabet also occurred; but after the eastern part of Moldavia was annexed as Moldova by the Soviet Union in 1941, Cyrillic was reintroduced. In 1989, though, the country reverted to the Latin alphabet, two years before it declared its independence. The interesting fact for the purposes of the present discussion is that the Cyrillic alphabet is better suited for representing Romanian and Moldovan than the Latin alphabet because of its larger number of letters, allowing it to represent all the sounds of the language individually. However, ideological desire (in the nineteenth century) to showcase descent from Latin and (in the late twentieth) to assert independence from the Soviet Union, together with the Moldovans’ wish to reunite the writing system with that of their Romanian kin, rendered any relevant linguistic or functional concerns utterly inconsequential.

It is, moreover, far from clear in any case that a precise phonetic script is always the ‘ideal’ writing system. Mandarin Chinese illustrates this issue very clearly. On the one hand, the traditional logographic writing system is highly complex, requiring the mastery of several thousand characters for competent literacy. On the other hand, the language has an extremely high rate of homophony due to a combination of two factors: first, most morphemes are monosyllabic, and second, the number of distinct syllables is small, due to historical loss of most final consonants and reduction or simplification of initial complex onsets and various consonantal contrasts.Footnote ²⁴ Thus an alphabetic rendering of a syllable such as yì, while simple to learn in the abstract, would be horrendously ambiguous (especially outside of context) since this syllable corresponds to over thirty homophonous morphemes,Footnote ²⁵ but when each homophone is represented by a different character, the ambiguity is erased. A similar case can be made for preserving the heterography of English homophones like air, heir, ere and e’er. On these grounds alone it is not mysterious why the various ways of writing Chinese phonetically that have been invented over the centuries (whether based on traditional characters, like bopomofo or zhùyīn, or using the Latin alphabet, like pīnyīn) have never been adopted for general employment. Probably the most widespread use to which such phonetic systems have been put is as so-called ‘ruby characters’ written above or to the right of a character to indicate pronunciation in children’s books; Japanese children’s books use kana for the same purpose to indicate the pronunciation of kanji. Chinese characters also have an important unifying role at the cultural and political level, as they are used throughout the country as the standard written medium across regions with mutually unintelligible local varieties of Chinese.Footnote ²⁶

9.5 Subphonemic Indication

Phonemes are pronounced differently in different contexts; these variants are called allophones and comprise the subphonemic level of speech. In relaxed speech in most varieties of English, the /t/ of top is pronounced quite differently from the /t/ of butter, and also (if a bit more subtly) differently from the t of stop and tray. The occurrence of a given allophone is always predictable based on the surrounding context. Subphonemic distinctions such as these are generally ignored in writing systems (native speakers may not even be aware of them). A notable exception is Sanskrit (Reference WhitneyWhitney 1889). The intellectual life of ancient India was focused on grammar, whose analysis was developed with a sophistication unique in the ancient and medieval worlds. This left an indelible stamp on the orthography of Sanskrit, which encodes an unusual amount of subphonemic detail. In particular, the phonetic changes that happen at word junctures (external sandhi rules) are clearly indicated, in contrast to the orthographies of other languages, where such aspects of pronunciation are routinely ignored. Thus the English verb hit is so written regardless of whether, in relaxed speech, one pronounces it as /hɪʔ/ with a glottal stop (as in he hit me), /hɪɾ/ with a flap (as in I hit him), or /hɪtʃ/ (as in he hit you) where the /t/ coalesces with the following /j/ to form a /tʃ/-sound. Notably, Sanskrit reflects such variants in its spelling. Hence, for example, the ablative singular form nagarāt ‘from (the) city’ undergoes changes in spelling in combinations such as the following (due to the nature of the script, which is an abugida or alphasyllabary, words ending in a consonant are written together with the following word): nagarāt + agan ‘(s)he went’ > nagarādagan, nagarāt + śukrāt ‘shining’ > nagarācchukrāt, nagarāt + jagāma ‘I went’ > nagarājjagāma, nagarāt + navyāt ‘new’ > nagarānnavyāt.

Though this system is carefully followed in all word-combinations in modern editions of Sanskrit texts and is described in detail by the Sanskrit grammarians, fidelity to it varies widely in manuscripts and inscriptions. In the prenormative language of the oldest Sanskrit text, the Rigveda,Footnote ²⁷ the external sandhi rules are not always the same as in the Classical language and are not applied in a number of circumstances. In Middle Indic, sandhi is only indicated in word-groups that are closely connected syntactically. All this suggests that the rigid, across-the-board orthographic representation of external sandhi is artificial and not reflective of the reality obtaining in speech.Footnote ²⁸ More normal in the ancient world is sporadic indication of external sandhi, typically in close syntactic contexts like short prepositional phrases. This is found frequently, for example, in Greek and Latin inscriptions.

Elsewhere, indication of subphonemic detail is sporadic at best. Probably the sandhi phenomena that are most consistently observed in spelling systems are elision and contraction, the former of which in particular can be mandatory even at the slowest talking speeds. Thus French l’ami ‘the friend’ is never written le ami because it is never so pronounced (versus, e.g., in the placename Le Havre). One of the earliest Greek alphabetic inscriptions, on the so-called Nestor’s Cup from around the late eighth century BC (see Reference WatkinsWatkins 1976), indicates elision in the sequence (with restorations) ΗΟΣΔΑΝΤΟΔΕΠΙΕΣΙΠΟΤΕΡΙΟ hosdantōdepiēsipotēriō = hos d’ an tōde piēsi potēriō ‘but he who drinks from this cup …’, where the adversative conjunction de is elided to d’ before the vowel-initial modal particle an. Elision in French is regularly indicated as well; this is true also of the opposite phenomenon, liaison, an external sandhi phenomenon whereby etymological word-final consonants that are not ordinarily pronounced come out of hiding (as it were) before a vowel in certain syntactic contexts. Liaison is normally indicated by default since the majority of such consonants are already present in the spelling for historical reasons but are otherwise silent (due to their having been lost through sound change over the centuries); in certain cases, however, the consonants do not normally occur in spelling but are restored in liaison contexts, for example, elle parle ‘she speaks’ but Parle-t-elle? ‘Does she speak?’, donne-le-moi ‘Give it to me!’ but donnes-en ‘Give some of it!’, with restoration of the historical verb personal endings -t and -s that dropped out of pronunciation in all other contexts.

A common subphonemic rule of many languages is the devoicing of word-final voiced consonants. In some written traditions, such as those of the Slavic languages, this devoicing is not reflected in spelling: thus Czech hradu /ɦradu/ ‘of castle’ (genitive singular) but nominative singular hrad even though pronounced /ɦrat/. Others, nonetheless, do spell out the alternation, such as Middle High German, for example, dative singular lobe ‘fame’ but nominative lop. This practice, however, died out; in present-day German, spelling is morphophonemic, whence nominative singular Lob even though the final b is devoiced.Footnote ²⁹ In all the examples given so far where a spelling system represents a subphonemic distinction, the allophones in question nevertheless are the same as other phonemes that are already represented by graphemes. Thus in the Middle High German example, lop is written with a p because a b in that position sounded like a p, which is itself a phoneme (a different phoneme from b, of course). Many allophones, however, do not sound like phonemes elsewhere in the system, as is the case with several of the English examples: the t of matter (in the variety of English described here) is a flap, which is not a phonemic sound, and likewise the t of hit (a glottal stop). It is very rare for writing systems to create dedicated symbols for allophones of this type. Once again, Sanskrit provides a couple of examples, though usage is not consistent at all historical periods and in all varieties of the Brāhmī abugida; one such example is the symbol ḷ, which stands for an allophone of ḍ between vowels.

9.6 Beyond the Level of the Grapheme and Phoneme

To those used to the fixity of spelling in our contemporary orthographic world, where usually only one way of spelling a word is prescriptively ‘correct’, the looser practices that obtained throughout most of the rest of human history can seem puzzling and disorienting. One of many examples that could be cited comes from Hittite. Though the cuneiform writing system was perfectly capable of indicating contrasts in vowel length (pa versus pa-a) and single versus geminate consonantism (a-pa versus ap-pa), it was not part of the scribal culture to indicate such distinctions consistently except in certain words. There is consequently much fluctuation in spelling. The variety only increases when another factor is added, namely, that the distinction between voiced and voiceless stops that the syllabary actually encodes was not in fact implemented by the Hittites when they adopted the script, and they used the corresponding symbols interchangeably. Thus the stem of the noun for ‘body, self’, which was phonetically [tweːk-], shows up in no less than five different spellings, as witnessed by fluctuations in spelling like tu-eg-ga-az, tu-e-eg-ga-az, du-eg-ga-az, tu-e-ek-ki and tu-ek-ku-uš (the different endings reflect different grammatical forms of the word).Footnote ³⁰

Of course contemporary orthographic systems have their inconsistencies as well, but these tend to be standardized for individual lexemes rather than the result of free variation/choice. The same symbol or orthographic sequence can have several different pronunciations (as the sequence ough in English though, through, rough, cough, thought, bough, plus the British hiccough) and the same phonetic entity can have several different ways of being represented in spelling (as with long /aː/ in the German words Tal, Wahl, Aal, versus only one way to indicate short a in the same environment: Fall ); but the spelling of each word is fixed in these traditions. Sometimes having more than one spelling for the same phonetic sequence can serve to disambiguate. Thus in German, the diphthong /aj/ is normally spelled <ei˃, but <ai˃ is also used, especially for the less frequent of two homonyms (e.g. Weise ‘way, manner’ vs. Waise ‘orphan’, Seite ‘page’ vs. Saite ‘string of an instrument’). A present-day language where spelling still has a bit of freedom is Japanese, though the freedom operates within a fairly limited set of choices. As briefly discussed in Chapter 7, three separate writing systems are used, kanji (logographic Chinese characters representing lexical sememes) and the two syllabaries (hiragana and katakana) collectively called kana that have already been mentioned and that are purely phonetic. While one could theoretically use kana to write an entire text phonetically, this is rarely done: kanji and kana are used side by side, with the root meaning of full lexical words typically expressed with kanji and stem and inflectional endings added using kana. The latter, nevertheless, can always be used instead of kanji depending on authorial choice and stylistic factors; broadly speaking, the more kanji that are used, the more formal the style. The fixity of orthographies over much of the present-day world is a product only of the last very few centuries, often the result of the workings of a centralized normative institution or academy, but sometimes (as in English) the combined result of efforts by individual eminent literary figures who took it upon themselves to promulgate orthoepic norms. In parts of the ancient world where alphabets were used and where there was close to a one-to-one correspondence between letter and sound, one sometimes sees a high degree of regularity in spelling (e.g. in Classical Attic Greek inscriptions); but it is not clear if the regularity resulted by default from the general accuracy of the script, or whether it was purposefully imposed.

Many times in the history of writing, cultural movements arose resulting in broad spelling reforms through imitation of the orthography of a prestige language. Throughout the last two millennia of Western European history, Latin and Greek have enjoyed such prestige status and have influenced the spelling (and vocabulary) of most Western European languages to a sometimes remarkable degree. A particular wave of such respellings gripped these literary cultures during the Enlightenment in the seventeenth century. Reference BrengelmanBrengelman (1980) regards the central impact of this intellectual movement on spelling to have been rationalization or regularization, with the result that, for English at least, by the 1660s the plethora of earlier spelling variants had largely disappeared and a widely agreed-upon system was in place that has remained mostly unaltered to the present day. Words of Latinate origin underwent consistent respelling according to their etymologies, many of which resulted in changes to pronunciation on the basis of the new spellings (e.g. adventure, earlier aventure), others of which did not (e.g. debt, earlier dette, compare Lat. debitum).Footnote ³¹ In French, there were parallel efforts. The French descendant of Latin grandis ‘large’ was spelled grant (masculine) and grande (feminine) in Old French, which was phonologically accurate (the dental had undergone final devoicing in the masculine); but the masculine was later respelled as grand, even though the historically correct pronunciation with /t/ persists to this day in liaison contexts (e.g. grand homme ‘great man’ /gʁɑ̃t‿ɔm/). As these examples show, such reforms do not result in greater regularity or systematicity across the board, and can actually increase the divide between spoken and written language.

9.7 Conclusion

In spite of being perforce cursory, the above overview has attempted to showcase the great variety of factors – many of them linguistic but just as many of them not – that can either advance or impede the achievement of systematicity in writing and orthographic systems. Given the number of different modes of pronunciation that obtain even within a single speech community and given humans’ great ability to make do with and make the best of imperfect conditions and tools, it is not surprising that so few (if any) systems of writing or orthography are fully consistent at any level of abstraction.

It may have been noticed that the words ‘systematicity’ and ‘idiosyncrasy’ have not been given strict definitions in this chapter beyond anything that their ordinary usages denote. This was done on purpose since, at least to an extent, whether something can be called systematic or idiosyncratic is in the eye of the beholder and depends on factors like the level of resolution and the context: what might seem systematic at one level or in one context can appear an idiosyncrasy in another. Not only do writing systems exist on many planes, they are also not merely the products of individual intellects, nor merely visual symbolizations of structures of spoken language, but expressions of culture at various aesthetic and functional levels. Exactly how the connections between writing and culture are to be theoretically modeled is a very complex one and is far beyond our scope; other scholars may find it useful to explore this area against the backdrop of more general anthropological investigations of specific cultures or the human species more broadly. Of course it need not be the case that each and every discrete aspect of a writing system is culturally underpinned, and the same goes for any specific instance of systematicity in writing or its absence.

Regardless of the resolution of this question, its very existence creates challenges for another issue that has not been considered in this chapter, namely, whether any useful comparative typology of idiosyncrasies can be catalogued. It would appear to emerge (at least preliminarily) from the foregoing that such a typology, however descriptively or conceptually useful, would nonetheless be of limited intellectual or explanatory value, focusing as it would on epiphenomena. Systematicity is an imposed structure or framework that is deployed to varying degrees by the creator(s) of a given writing system; idiosyncrasy arises (or can arise) wherever this framework is not applied. The motivations for its application or nonapplication are really of greater interest, assuming they are recoverable from the historical record or indirectly through other historical and cultural investigations.

10.1 Introduction

When exasperated language purists post online indignant comments on ‘the current state of the language’ they frequently pin the imminent demise of ‘proper language’ on abbreviations. Young people in particular, they claim, obsessively and compulsively truncate every other word when they engage in written interactions, manifesting their ‘sloppiness’ and ‘laziness’ in utter disregard for the sanctity of the dictionary form. Yet those very same people who lament the abbreviative nature of modern communication will roam the streets of Rome or take a trip around Cyprus with nothing but admiration for the monumental architecture, adorned with inscriptions, of which abbreviation is a common element. Still, despite their ubiquity, signs of truncation and curtailment would prove problematic even for the students of diachronic orthographies (with the notable exception of paleographers) who did not quite know how to incorporate scribal abbreviations into linguistic analyses of historical texts (see, e.g., A Linguistic Atlas of Late Mediaeval English, Reference 725McIntosh, Samuels, Benskin, Laing and WilliamsonMcIntosh et al. 1986a). Indeed, truncating frequently recurring, familiar items has always been part of the written culture. Limitations of space, workability of the writing surface or the time necessary to do the job all required the development of ‘coping strategies’ from the writers, whether they worked with stone or parchment. Rather than marking any ‘impoverishment’ of the language, abbreviations were a direct appeal to readers’ literacy competence, which entailed a knowledge of sometimes intricate but mostly stable mappings between these visual symbols and their orthographic interpretation. Indeed, to abbreviate (and to understand what abbreviation stood for) meant to be a litteratus and, in a way, this still holds true in the modern digitally oriented written communication.

This chapter, while limited in its scope, is thus a reminder of the longue durée of abbreviation in scribal practices of the early-to-late Middle Ages in parts of Western Christendom, specifically England and Ireland. The Anglo-Irish context is both typical and idiosyncratic in the ways in which abbreviations were integrated into literacy practices of textual communities in the medieval and early modern periods. On the one hand, Irish and English scribes would access the same pool of symbols with the same range of meanings that were available for all writers and readers in the Latin West. Barring regional modifications to a handful of symbol shapes, the Latin system of abbreviations transgressed geographic and political boundaries in times when vernacular languages were not the default option for all written works, and when, regardless of their mother tongue, scribes from all over medieval Europe were practitioners of a form of Latin-based literacy. On the other hand, both geographic and, in a sense, cultural distance from continental influences (including a very distinct path of Christianization of the British Isles) facilitated a distinct ‘vernacularization’ of insular abbreviations.

By tracing the origins of abbreviations in late Antiquity and recounting their spectacular success in the Middle Ages, followed by their gradual (yet never complete) fall from grace in the early modern period, this chapter points to the lasting impact of abbreviation on contemporary abbreviating habits in written communication. Not unlike their contemporary manifestations, the abbreviations which are the focus of this chapter were always ‘shape-shifters’: not only would they change their form depending on the type of script, position on the page, or the language of the text but also they were ever negotiating between logos and imago, text and image, shape (figura) and its phonetic interpretation (potestas). In that sense, they serve as links between the different modalities which informed premodern textuality, entreating the reader always to shift between the linguistic, visual and spatial elements of the manuscript (or early printed) page to make sense of the text before their eyes. Medieval abbreviations can therefore help us to better understand the multimodality which informs modern, digitally based communication. Abbreviations are not and never have been foretellers of some kind of linguistic apocalypse; if anything, they reflect the multilayeredness and multiaspectuality of visually based transmission of ideas.

Arguably, one of the factors contributing to the ‘problematic’ reputation of abbreviations is the fact that these symbols span the boundary between the linguistic, visual and spatial components of the manuscript page. As such, they are neither the focus of linguistics sensu stricto, nor do they belong with visual studies. Moreover, the near-identicality of forms for Latin and vernacular languages suggests the commonality of functions in both linguistic contexts, even though they demonstrate patterns of distribution more complex that what could be explained away by the different structures of Latin and vernacular orthographies. The “half-graphic” structures – as Reference TraubeTraube (1907) once referred to them – sit inconveniently between image and text, the visual and the linguistic, the figura and the potestas (i.e. shape and sound value), Latin and the vernacular. Historically, abbreviations were applied to frequently recurring grammatical items (including prepositions, particles or inflectional endings) and for words which either could not be spelled in full out of reverence for the deity (nomina sacra) or needed not be spelled out because they belonged to a specialist jargon (notae iuris). However, in translation from Latin to vernacular languages those contexts and connotations were largely lost and what remained was their orthographic similarity (or identicality) with particular vernacular forms. Thus, from the linguistic point of view, the function of abbreviations shifted from that of grammatical or lexical markers to shortcuts for a specific orthographic sequence, which might signal a specific morpho-lexical category.

At the same time, what is becoming increasingly clear is that, far from being just a scribal shorthand, symbols of abbreviation are a crucial element of medieval multimodality (see Reference AmslerAmsler 2012, Reference Amsler2016, Reference 724MaxwellMaxwell 2016). That is to say, they often participate in the creation of meaning of the text alongside other semiotic resources: it is not just the linguistic layer but also the visual makeup of the manuscript page (and the interactions between such visual elements as ink colors, script sizes, letter shapes), the spatial arrangement of its components or the tactile experience of the page itself all add to the pragmatic reading of the text. It is also through multimodality that one can access the multilayered functions of abbreviations, more about which is said in the subsequent parts of the chapter.

10.2 Polyvalent Functions of Abbreviations

“Take a foreign language, write it in an unfamiliar script, abbreviating every third word, and you have the compound puzzle that is the medieval Latin manuscript” – with these words Reference Heimann, Kay and CapellliHeimann and Kay (1982: i) begin their translation of Adriano Cappelli’s Dizionario di abbreviature latine ed italiane (Reference Cappelli1899), which is to date one of the key references for those who work with medieval manuscripts. Indeed, so critical are abbreviations to pre-print textuality and so common are they in scribal output that the disciplines of paleography and diplomatic(s) (invested in the study of old handwriting and critical analysis of historical documents, respectively) trace their late seventeenth-century/early eighteenth-century origins to the very practical need to decipher symbols with which medieval scribes interspersed their writing. In the most basic definition, abbreviation is the shortening of a word either by disposing of some letters from the end (or from the body and the end) or by curtailing the letters from the middle but leaving the initial and final letters intact (Reference ThompsonThompson 1893: 75). While individual abbreviation shapes would undergo mild-to-serious modifications and the number of types belonging to active scribal repertoires would fluctuate from one period and one geographical region to another, their primary purpose has always been to facilitate the efficacy of writing; either by truncating the familiar and recurrent orthographic strings and thus speeding up the copying process or by pointing to the nontextual reality (e.g. through legal shorthandFootnote ¹ or nomina sacra).

Abbreviations are peculiar to pre-printed texts also because they are a link between text and image: in the highly visual literacy of the Middle Ages they are the forms combining littera with imago. Now based on individual letter shapes, sometimes modified by additional strokes, now entirely ‘un-letter like’ and arbitrary, symbols of abbreviation sit somewhat uncomfortably between the visual and the textual. Hence, no wonder that they should have been approached from either of the two perspectives (visual, i.e. paleographic, or textual, i.e. linguistic) rather than a combination of the two. The typical manner in which medieval abbreviations have been presented to the modern reader is in the form of abbreviation dictionaries, the most influential of which were composed in the eighteenth and nineteenth centuries. Apart from Cappelli, mentioned above, one should also mention Reference ChassantChassant (1846) and Reference Trice MartinTrice Martin (1892), both of whose dictionaries problematize Latin abbreviations in the context of vernacular medieval sources (French and English, respectively).

While extremely useful for the modern reader of medieval sources, these dictionaries, with their impressive lists of Latin abbreviations, give an appearance of stasis, which belies the actual reality of the early-to-late Middle Ages. One would be arguably hard-pressed to gather from the lists collated by Chassant, Cappelli or Trice Martin that these scribal (and, to a certain degree, typographical) symbols were actually a context-bound, dynamic and open category, reflecting the linguistic, cultural and very pragmatic needs of the people who applied them to parchment or paper. Just as the Latin of the Roman republic (with which period the increasingly widespread use of abbreviations is typically associated) bears little semblance to the Latin of the medieval Church, so the system of abbreviations used in the legal documents of late Antiquity is different from that of religious texts of the high-to-late Middle Ages. Apart from historical depth, there is also diatopic differentiation in the abbreviation systems applied throughout our period of interest: even before the vernacularization of writing scriptoria in different parts of medieval Europe (and often, too, different scriptoria within the same political unit) there were specific in-house practices for abbreviations. Of course, in the era of Latin as a written lingua franca, there needed to be limits to the inventiveness of the scribes and an adherence to a general set of rules was necessary, lest written communication collapse. There are peaks and troughs in the history of abbreviations: the ‘need for speed’ ultimately had to be reconciled with practicality, and the scribes needed to know how to strike the balance between efficient copying and readability of their texts. Up until the twelfth century AD, the number of abbreviations was on the rise, with the Carolingian Renaissance of the ninth and tenth centuries contributing significantly to the expansion of the system and creation of new forms (Reference 676DerolezDerolez 2003: 72). When Latin lost the hegemony of being the written medium of the educated litterati, abbreviation systems started to be regulated by vernacular orthographies. A number of forms (like ꝝ for -rum, bȝ for -bus, qꝫ for -quae abbreviations) had to go, simply because vernacular languages did not have appropriate referents for them.

At the same time, scribes operating in vernacular systems devised new abbreviations for items not encountered in Latin (Irish scribes are said to have been particularly fond of that practice; see Reference BischoffBischoff 1990: 91). For late medieval and early modern scribes, abbreviation was as much a matter of the economy of copying as it was a marker of the formulaicity of written language: some types of texts and some formats of manuscripts are just more likely to contain more abbreviations in general or to contain abbreviations of specific type(s). For instance, the most densely abbreviated manuscripts in the Middle Ages are those which were intended for study, whereas those meant to be recited from – like liturgical manuscripts – contain relatively few abbreviations (Reference Clemens and GrahamClemens and Graham 2007: 89). As such (and as has been argued elsewhere; see Reference Rogos-HebdaRogos-Hebda 2016, Reference Rogos-Hebda2020) they too, along with elements of the visual makeup of the manuscript page, indicate the pragmatic functions of the text, rather than simply encode specific orthographic strings to save the scribe’s time. This richness of forms and functions stems from centuries-long processes of adaptations and modifications, which still resonate in the postprint world of today. This chapter traces the origins of symbols of abbreviation in the Latin West and outlines linguistic and cultural contexts for their changing graphic forms from late Antiquity to late Middle Ages (first century BC to fifteenth century AD). Section 10.3 focuses on sources for abbreviations in the written culture of the Latin West. Section 10.4 discusses typologies of abbreviation systems, circulating in manuscript studies, paleography and codicology and characterizes each of the types. Section 10.5 looks into modifications (or, better perhaps, adaptations) of the system inherited from late Roman practices and introduced by scribes translating that system into vernacular forms. Moreover, a link between the types of script and the type of abbreviation forms is made explicit. Finally, Section 10.6 hints at the transition period between the gradual decline of handwritten book production and the burgeoning print market, signaling some of the printers’ responses to the abbreviations.

10.3 Sources for Abbreviations: Sigla, Notae Iuris and Nomina Sacra

All writing is abbreviative, to a degree. Limitations of the writing surface, pressure of time and the scribe’s own predilections have always featured in the decisions to implement abbreviation. Some textual contexts would entail abbreviation as a matter of course, as students of numismatics and epigraphy will testify. Words encarved in metal and stone would be abbreviated for practical reasons (the availability of space), but they would also serve as visually encoded markers of power structures (coin issuance and architectural commissions were the prerogatives of the powerful). Yet it was inscriptions that, despite their abbreviative character, were able to communicate longer portions of text, thus becoming a template for the practices of scribes engaged with more ‘workable’ materials. In paleographic literature, the origins of what became the Latin-based system of abbreviations, used in the Latin West between late Antiquity and the early modern period (approximately seventh to sixteenth century), are traced to the sigla system and the Tironian notes of the Roman Republic on the one hand (Reference CappelliCappelli 1899: 1) and, on the other, to the nomina sacra of the early Greek manuscripts of the Scriptures (Reference TraubeTraube 1907: 45).

A siglum was form of abbreviation in which a word was truncated to its initial letter (as in A.D., S.P.Q.R.). Common in inscriptions, they gradually made their way to shorthand, which was becoming more widespread in the late Republic (approximately first century BC). These so-called litterae singulares were commonly applied to first names, calendar signs or common formulae in legal and business writings of the Classical Latin period (i.e. first century BC to third century AD) (Reference BischoffBischoff 1990: 150). Reference CappelliCappelli (1899: 3) refers to sigla as the most important but also the most notorious type of abbreviation by means of truncation (where the only unabbreviated part of a word is its initial part), which could only be efficient so long as it was limited to the most frequent lexemes. Sigla usually took the form of a majuscule initial followed by a period, although in written documents it is also possible to find sigla written as minuscules, often not followed by a period but marked by a horizontal stroke above the siglum (Reference CappelliCappelli 1899: 4). Moreover, in medieval written documents, one can find sigla made by the middle rather than initial letter of a word (e.g. h for nihil ). Doubling the majuscule, in turn, would indicate plurality or the superlative degree (e.g. FF. would stand for fratres). For female first names and titles, sigla would have been tilted or written upside down (e.g. W. for Mulier). Reference CappelliCappelli (1899: 6) links the gradual decline of the sigla system to its inherent inefficacy: there are only so many words which can be unambiguously assigned to this category, and the eighth and ninth centuries witnessed the appearance of abbreviations which truncated fewer elements than sigla did and thus improved the legibility of such abbreviations.

From truncating words to just the initial letters developed the practice of leaving either the first letter of every syllable in a word unabbreviated, or else of leaving the first and final letter. This principle of contraction was applied in notae iuris – a system of legal abbreviations which, albeit more systematic than sigla ever were, were often ambiguous (Reference Di RenzoDi Renzo 2000: 7). Lack of punctuation and adequate spacing in writing, combined with the propensity for abbreviating legal formulaic language, which was of little use and even less practicality beyond the legal context, made notae iuris hardly amenable for use in more general contexts.Footnote ² That much was apparently noticed by Marcus Tullius Tiro, the secretary of Cicero, who needed a system for recording his master’s dictations, public speeches and business transactions efficiently. His system, which came to be known as notae Tironianae, relied on three pillars: Latin letters, elements of Greek shorthand, and abstract symbols devised by Tiro himself (Reference Di RenzoDi Renzo 2000: 8). A distinguishing feature of Tiro’s system, and one of the causes of its success and subsequent expansion, was that, apart from developing abbreviations for individual words, it also took care of broader grammatical categories. Thus, Tironian notes included symbols for recurring inflectional endings, prefixes or function words (viz. abbreviations for et, enim, esse, sunt ). Combinations of abbreviated letters, syllables and entire words then were used as shorthand for recording whole sentences and thus became the most varied and accurate shorthand system in the Antiquity (Reference Di RenzoDi Renzo 2000: 8).

The system of ‘slashes, curves and hooks’ (Reference TeeuwenTeeuwen 2014) would survive until the Middle Ages, when it was used predominantly in the context of legal and administrative writing. The system of Tironian notes would be further expanded and, from the early medieval period onward, combined with other abbreviations (Reference RussonRusson 2016). The original system swelled to more than a dozen thousand symbols by the beginning of the Carolingian period (mid-eighth to ninth century), after which the numbers still in use were significantly smaller. Reference TeeuwenTeeuwen (2014) highlights the fact that the presence of Tironian notes in a medieval manuscript is indicative of the scribe’s high level of education; an education most likely received in a monastic institution invested in scholarly activity. While one can tease out individual symbols for particular abbreviated items (lexical and/or functional) from the extant sources, it can be frustratingly difficult to interpret actual Tironian notes in a medieval text. The basic elements (i.e. strokes and curves which make up the notae) can be connected with following figurae (i.e. letter shapes) in highly idiosyncratic manners. Simply put, it is a challenging task to decipher the Tironian shorthand of a writer whose hand one is not familiar with (see Reference TeeuwenTeeuwen 2014). With time, the number of Tironian notes used by the scribes was significantly limited, as limited was the audience able to read this shorthand. It is with a scholarly setting that one should associate this system of abbreviations and, apart from a brief period of rekindled interest in the system during the twelfth century, it mostly went out of use, save for a few remnant symbols, which survived up until modern times (e.g. the Tironian et : in Irish and Scottish signs).

The third source for medieval abbreviations are the early Greek translations of the Septuagint or, more specifically, the small set of abbreviations for terms referring to the names and attributes of God. Reference ThompsonThompson (1893: 76) connects the emergence of the nomina sacra, as these abbreviations came to be known, with the practice of Hellenistic Jews, working on the translation of the Septuagint from Hebrew into Greek. To illustrate, rather than spell out the name Yahveh, these scribes tended to either copy the Hebrew Tetragram YHVH (the vowel-less version of the name of God) into the Greek text or, similarly, omit vowels in the Greek ƟEOC to ƟC, out of reverence to the deity. Along the same lines, KYPIOC for ‘Lord’ would become KC and both abbreviations would receive a horizontal stroke above. In the Hebrew tradition of copying, using such strokes would typically single out specific words as foreign or emphatic, but since a horizontal line was applied to abbreviated forms of ‘God’ and ‘the Lord’, it came to be used as a general symbol of abbreviation and would extend over the entire abbreviated word (Reference ThompsonThompson 1893: 77). Initially, the group of sacred names (the nomina sacra) was limited only to five nomina divina, ‘godly names’, namely God, the Lord, Jesus, Christ and Son. It later expanded to 15 items, including words for Spirit, father, savior, heaven, man, Israel, Jerusalem, David, cross and mother (the latter only appearing as nomen sacrum in the fourth century AD).

Unlike notae iuris or notae communae, nomina sacra were not motivated by considerations of space or the speed of copying. The scribes’ decision to resort to contracted forms of the 15 terms listed above was one guided by reverence rather than economy. The abbreviating principle with this category of lexemes was one of contraction, that is, abbreviating the middle of a word, leaving the first and final letters unabbreviated (and thus allowing an easy identification of case ending). While in principle biblical manuscripts precluded abbreviation (so as not to lead to interpretative ambiguity), by the Roman period “abbreviations of terms central to Christian worship” (Reference BischoffBischoff 1990: 152) were common enough. In Latin copying tradition, Greek nomina sacra were transliterated, with one notable exception, namely XPC, which often retains its Greek lettering in Latin texts, albeit receiving Latin inflections (Reference Horsley and WaterhouseHorsley and Waterhouse 1984: 211). Thus, Deus, Iesus, Christus, spiritus, dominus (noster), sanctus (and later also David, Israel, Ierusalem) become, respectively, DS, IHS, XPS, SPS, DMS (N) and SCS. Toward the early Middle Ages (sixth to tenth century) a number of abbreviations which originated as nomina sacra came to be used with nonreverent meanings, for example, forms like DMS or SPS for dominus and spiritus, respectively, were largely restricted to their everyday connotations (Reference BischoffBischoff 1990: 152). With the expansion of the institutional Catholic Church in the high Middle Ages, the same principle of contraction resulted in the coinage of new abbreviations, denoting names of Church officials (episcopus and presbyter abbreviated as EPS and PRB are cases in point) (Reference BischoffBischoff 1990: 153).

10.4 Typologies of Abbreviations

Reading historical handwriting is largely an act of decoding: the unfamiliar scripts, the idiosyncrasies of letter shapes, the nonalphabetic symbols, some of which are ornamental flourishes while others abbreviate specific orthographic sequences; all these elements require a well-trained eye. The study, description and interpretation of historical handwritten material are the tasks of paleography, a discipline which, from its inception in late seventeenth century, would try to embrace the vast and inherently variable documentary corpus of late antiquarian and medieval texts under a comprehensive nomenclature. However, paleographic terminology has always been somewhat fluid: descriptive terms used for characterizing letter shapes and symbols encountered in a medieval manuscript can vary in detail among paleographers, although a common core of typology is generally accepted. Specifically, terms like lobe, shaft, bowl, arm, ascender, descender, etc. to describe elements of letter figurae recur in publications on the topic, although authors differ in the manner in which they describe the scribal hands or the ways in which they characterize pen lifts. Similarly, typologies of abbreviation symbols have differed slightly from one author to another, depending on the criterion adopted for organizing the system. Reference BrownBrown (1990: 1), in fact, insists that terminology is one of the most controversial (although also the most basic) issues in paleography. One cause for this is that a number of such categorizations are based on mixed criteria, for example: context/text type appears next to shape (viz. nomina sacra vs. special symbols ); mode of abbreviation next to ‘whatever does not fit’ (suspensions vs. contractions vs. brevigraphs); position within the line next to orthographic context (superior letters vs. signs significant in context). Even though abbreviation symbols feature in the works of both ‘founding fathers’ of paleography and diplomatics (the study of documents, as opposed to, e.g., literary texts), namely Reference MabillonMabillon (1681a) and Reference MontfauconMontfaucon (1708), it was not until the nineteenth century that some sort of taxonomy for abbreviations as a distinct category of scribal output began emerging (a historical outline of these taxonomies can be found in Reference Honkapohja, Meurman-Solin and TyrkköHonkapohja 2013). The most comprehensive and most general categorization, followed by many contemporary handbooks, is the one comprising suspensions, contractions and special signs (also termed ‘special symbols’ or brevigraphs). This type of classification combines the categories of ‘mode of abbreviation’ with the ‘physical appearance of the symbol’ as classificatory criteria. The most straightforward typology, reiterated in many publications, introduces the categories of suspensions, contractions and some sort of ‘special symbols’ type. While specific authors differ in their descriptions of details, general similarities in this tripartite division of abbreviation symbols can be observed.

Suspensions are those abbreviations which truncate the end of a word; they appear under this name in Reference ChassantChassant (1846), whereas Reference CappelliCappelli (1899: 1) refers to this category as “abbreviation by truncation.” An example of this abbreviation type are Roman sigla (e.g. D. for dux/dominus, J.C. for iuris consultus, Reference CappelliCappelli 1899: 4), but usually more than just the initial letter is left unabbreviated in this category (e.g. BO.ME. for Bonae Memoriae, Reference CappelliCappelli 1899: 5). Often, suspension is indicated by means of a punctus, following the letter which was left unabbreviated (Cappelli’s example is .n. for enim ; (Reference CappelliCappelli 1899: 4)). Reference CappelliCappelli (1899: 1) further differentiates between general and specific signs used for suspension: the former only indicate that the word has been truncated, the latter that abbreviation concerns the ending of a word. An example of general signs is the horizontal stroke, which, positioned above the abbreviated word, may indicate syllabic suspension (Reference Clemens and GrahamClemens and Graham 2007: 89) or, when positioned above a word-final vowel letter, indicates the truncation of a following n or m (in fact, Reference CappelliCappelli (1899) treats abbreviation of final m and n as a distinct class). For example, the abbreviated form of amen in Figure 10.1 (Reference CappelliCappelli 1899: 6) is a case of syllabic suspension, whereas Figure 10.2 (Reference CappelliCappelli 1899: 14) abbreviates word-final m in aliam.

Abbreviation by contraction, in turn, depends on shortening the middle part of a word – it may be only one letter or more. Here, too, the symbol indicating this form of abbreviation is one of general abbreviation, that is, a horizontal (straight or curved) stroke placed above the letter(s) that precede(s) the abbreviated sequence. This abbreviation symbol, also known as ‘macron’, was by far the most versatile and the most common graphic symbol of abbreviation used in medieval manuscripts (see Figure 10.3).

Figure 10.1 Syllabic suspension

(Reference CappelliCappelli 1899: 6)

Figure 10.2 Abbreviation of final m

(Reference CappelliCappelli 1899: 14)

Figure 10.3 The macron symbol used for contraction

(Reference Trice MartinTrice Martin 1892: v)

That same symbol also happens to be the most common form of a word-final flourish, which can be easily mistaken for abbreviation. This ‘otiose’ pen stroke would be added to the final letter of a word for purely decorative purposes, a practice quite common among Middle English scribes (Reference 676DerolezDerolez 2003: 187) but potentially confusing for the readers. Otiose spellings in English vernacular manuscripts include forms like men̑ for men, sonn̑ for son, founden̑ for found or torn̑e for turn (all examples found in late fifteenth-century manuscript: London, British Library, Royal MS 18 D II, of John Lydgate’s Siege of Thebes). In fact, English vernacular manuscripts of the later Middle Ages are notorious for this potential for confusion.

In principle, no abbreviation symbol was invariable: its particular shape was always adjusted to the type of script selected by the scribe, the amount of interlinear space, the position of the abbreviated word on the page (within the ruled column vs. the margins) or the shape of the neighboring letter. The macron, however, was possibly the most variable symbol of abbreviation, both in terms of its forms and specific functions (linguistic and pragmatic). For one thing, the figura, or shape representing this particular abbreviation type, has the largest number of allographs available to scribes. Interestingly enough, one can see that variety within the confines of a single copying stint performed by a single scribe. Depending as much on scribal idiosyncratic preferences as on the type of script and the amount of space between the x-height (i.e. the body of the letter, minus ascenders and descenders, which extend toward the headline and baseline, respectively) and the headline, that horizontal stroke would appear as a short thick line (typical but not exclusive to more formal scripts, like littera textualis), a wavy line (both in textura-type and cursive scripts), a crescent-shaped symbol, with occasional punctus underneath (especially popular with cursive hands, see Reference PettiPetti 1977: 22) or a long hairline stroke (see Figure 10.4).

Figure 10.4 Some variants of the macron

(Reference CappelliCappelli 1899: 2)

With cursive scripts the macron abbreviation is often physically attached to the letter it precedes by means of a counter-clockwise stroke, exercised without lifting the pen. This is especially true of letters m and n (the latter often indistinguishable from the u littera), occasionally word-final a. A variant of that same type is a counter-clockwise stroke with a change in the direction of writing, likewise attached to word-final m’s and n’s. There does not seem to be much compelling evidence that these graphic variants had different functions, other than highlighting a scribe’s reactions to the dynamics of script and the word/line context. The complicating factor, rather, lies in their orthographic interpretation in vernacular manuscripts (especially English, but also French), when word-final stroke can be interpreted both as abbreviation for n and as an otiose flourish, that is, a penstroke devoid of any linguistic meaning (forms like governaūce for governaunce are clear enough, but that same word might well be spelled governauce in an English manuscript of the later Middle Ages).

Further complicating the distinctions between the categories of suspensions and contractions is the fact that both types can be operated by the same sets of symbols, the most common of which is the general abbreviation sign, mentioned above. Reference TraubeTraube (1907: 47) suggested that the difference between the two types lies in their distribution: while contractions typically occur with Christian Latin items, suspensions are more typical of Greco-Roman vocabulary, but this distinction is by no means categorical. The third category, superscript (or, superior) letters, is sometimes treated as a subtype of contractions (e.g. Reference PettiPetti 1977: 22), as it depends on abbreviating some elements from the middle of a word. As the name suggests, superscript letters are vowels or consonants raised above the x-height of the preceding letter. They typically truncate orthographic sequences comprising the raised letter and the letter r (thus abbreviating ar, ra, er, re, or, ro, ir, ri, ru, ur )Footnote ³ and with this function they are easily transferable from Latin to the vernacular (e.g. g^ace is a frequent spelling for grace in English literary manuscripts between the thirteenth and sixteenth centuries). In Latin manuscripts, superscript vowels have a fixed meaning (Reference CappelliCappelli 1899: 32) and operate more like contractions proper (e.g. a^a stands for anima or alia, oⁱ for omni, u^o for uno or vero). In English manuscripts, in turn, superscript e and sometimes u become near-abbreviations (Reference BrownBrown 1990: 11) in þ^e, þ^u for the(e) and thou respectively.

In less frequent cases, when the raised letter is a consonant, its interpretation depends on that particular consonant: it can abbreviate just two letters or entire syllables, so superscript consonants are much more context-dependent than vowels. For instance, a raised c, superscripted to another consonant, usually meant ec or iec (e.g. ob^cto for obiecto ; Reference CappelliCappelli 1899: 33). Superscript n, in turn, when preceded by a q, abbreviated quando, and superior letter t next to a consonant would represent it (e.g. m^tt^t for mittit ; Reference CappelliCappelli 1899: 33–34). Another recurring superscript consonant is r, which, when preceded by t, abbreviates ur (here the 2-shaped figura for r alternates with superscript u, in its ‘regular’ and ‘serrated line’ version), for example cēt^rio for centurio (Reference CappelliCappelli 1899: 33). In English, a raised t is a common abbreviation for two recurrent function words, with and that, which are often spelled w^t and þ^t, respectively. Occasionally, a superscript s appears line-finally, both in Latin and vernacular manuscripts.Footnote ⁴ Finally, special symbols, or brevigraphs, can either resemble one of the letters omitted or be arbitrary in shape (Reference PettiPetti 1977: 23). More recently referred to as simply “symbols” (Reference BrownBrown 1990: 5, Reference Clemens and GrahamClemens and Graham 2007: 89), they are a cover term for what most nineteenth- and early twentieth-century paleography handbooks saw as distinct categories: “signs significant in themselves,” also termed “signs significant in context” (Reference CappelliCappelli 1899: 18), “conventional symbols” (Reference CappelliCappelli 1899: 39), “abbreviative signs” or “special signs” (signes abréviatifs in Reference ChassantChassant 1846: 26). What binds all these symbols together are their arbitrary shapes: although some of them do resemble actual letters of the Latin alphabet, they are nearly always modified by some additional figura (e.g. the ‘p-bar’ or ‘p-loop’ abbreviations for per, par, por and for pro, respectively, which feature in medieval manuscripts as variants of ꝑ and ꝓ). Still others are ideograms with no relation to actual letters of the alphabet (e.g. the ⁹ abbreviation, standing in for con/com, the ₉ for us or the ꝭ for is/es/is).

Since a number of those special symbols represented orthographic sequences recurring not only in Latin but also in vernacular languages, many survived beyond the Classical period well into the late Middle Ages (fourteenth to fifteenth century), and some of them still feature in early printed books. Among them were the 9-like character for con/com, the supralinear hook-symbol indicating re/er, the loop for final is/ys/es, the p-bar, the p-loop, the serrated line/oc-version of a supralinear symbol for ra, the 2-like or a-like supralinear symbol for ur, or the 9-like symbol in the supralinear position for us (Reference PettiPetti 1977: 23–24; see Figure 10.5).

Figure 10.5 “Abbreviation marks significant in themselves”

(Reference CappelliCappelli 1899: 13)

Somewhat limited distribution became the fate of the ser-abbreviation , which can be encountered until well into the sixteenth-century in English manuscripts. The so-called ‘hook’ abbreviation, representing the er or re sequences (or, e, when attached to word-final r ), was nearly as multifaceted as the macron. Most typically, it would take the form of a counter-clockwise curl, either physically attached to the preceding letter or separate from it. In some variants it would be similar to punctus interrogativus, the ‘question mark’ (minus the dot); in others it would look not unlike the hairstroke variant of the macron (this form can occur in cursive scripts). Most typically the curl would have been small, although some scribes would elongate the stroke to reach a few letters to the left. Other abbreviations, such as ꝝ for rum, bꝫ for bus, qꝫ or ꝙ for quod, truncating Latin grammatical categories, were lost in transition to vernacular book production. Otherwise, although the grammatical referents of some of the abbreviations were gradually lost (like ꝰ and ꝭ, which in Latin represented inflectional endings), their orthographic interpretation remained available for vernacular languages too. Thus, in English the ꝰ and ꝭ brevigraphs remained in use to indicate plural ending or the possessive inflection. From the eleventh century onward, abbreviations became increasingly more common in manuscripts and new forms were introduced to cater to the needs of an expanding book market (an example of such innovation was the Tironian ⁊ for et ; Reference 676DerolezDerolez 2003: 66).

The peak of abbreviating practices, however, occurred in the thirteenth and fourteenth centuries, when the concurrently progressing vernacularization of book production and the emergence of new (i.e. lay, nonscholarly) readers marked the beginning of a gradual reduction of the numbers. Scholastic and scientific works, designed for limited audiences of professional litterati tended to retain their highly abbreviated nature, but works of literature and those produced for display (whose purposes would often overlap and coincide with public reading practices) were characterized by a significantly lower number of abbreviations, only limited to truncating the most common grammatical words and some conventionally abbreviated items. It is in those vernacular manuscripts that abbreviations acquire new, pragmatic functions, for example, when they signal code-switching along with other visual elements on the manuscript page (Reference Rogos-Hebda, Włodarczyk, Tyrkkö and AdamczykRogos-Hebda 2023) or when they aid the scribes in organizing the discourse on the page by combining with other visual discourse markers (see Reference Carroll, Peikola, Salmi, Varila, Skaffari and HiltunenCarroll et al. 2013, Reference Rogos-HebdaRogos-Hebda 2020).

10.5 Abbreviations and Medieval Scripts

Textbooks of paleography typically provide lists of Latin abbreviation symbols with their canonical orthographic interpretation(s), offering limited commentary on their diachronic and diatopic developments. While on the whole, the ‘meaning’ (in the sense of orthographic interpretation) of abbreviation symbols, which originated in Latin, remained unchanged throughout the Middle Ages and beyond, the morpho-lexical and visual contexts for those symbols did not remain unchallenged. Since every act of translating from one medium to another is inherently interpretative, whatever paleography textbooks, translating from manuscript to print, represent as ‘medieval abbreviation symbols’ is at once an approximation and a simplification of the visual-orthographic contents of the manuscript. Whether traced by hand from the actual manuscript (see Reference CappelliCappelli 1899) or printed in the so-called Record Type (a special type of font used up until mid-twentieth century to represent the minutiae of medieval manuscripts in near-facsimile editions, e.g. Reference Trice MartinTrice Martin, 1892), the symbol presented on the page to the contemporary reader is both an idealization and a ‘catch-all’ shape for the variety of forms available for one abbreviation. One source of that variety is the variation of post-Roman scripts themselves. Often referred to as ‘national hands’ (a term as anachronistic as it is misleading), scripts pre-dating the pan-European Carolingian minuscule were “fluid, interactive variants, in an unstable political climate” (Reference MarcosMarcos 2017: 22), each with their idiosyncratic approach to abbreviations. For example, the most successful (and longest-living) of these scripts, Insular minuscule (in its Anglo-Saxon minuscule variant), developed a series of contractions for the most frequently recurring elements of lexicon: and/ond ~ 7 (and its crossed variant ), æfter ~ æft̄, for ~ f̄, ge ~ g̅, þæt/þat/þet ~ ꝥ, sancte ~ sc̅e, þonne ~ þon̅, -um/-un ~ u̅ (Reference Rumble and RichardsRumble 1994: 14). Also, an Anglo-Saxon innovation was the tur abbreviation in the form of a “<t> with a downstroke through the bar” (Reference BischoffBischoff 1990: 91).

Even beyond the Carolingian period (ninth century), well into the reign of Gothic scripts, which leveled the differences between regional hands, one can notice geographically determined differences in abbreviating practices. Examples of such differences are distinct figurae of the bus-abbreviation in German-speaking and Scandinavian countries (b with a vertical zig-zag line instead of the more usual b followed by a semicolon-like character); idiosyncratic forms of con/cum abbreviations in Central Europe and Germany (inverted c formed with two strokes, rather than a 2-shaped, a 9-shaped or an inverted-c figura); and the Insular forms of enim- (n between two dots) and est-abbreviations (horizontal stroke with a dot or flourish above and below) (Reference 676DerolezDerolez 2003: 67). Individual scribes too had at their disposal more than one graphic variant of a given abbreviation and would sometimes apply different forms of the same abbreviation on the same page. Earlier textbooks would often attribute this variety to the “caprice or carelessness of the scribe” (Reference Trice MartinTrice Martin 1892: v), a judgment clearly colored by the standardizing effects of print, but inadequate to the context-driven motivations of scribes. The type of script as a geographical criterion has already been mentioned, but it needs to be emphasized that, beyond the ‘local flavor’ in abbreviating practices, scripts, whether transnational (however inadequate the ‘national’ attribute is when discussing premodern phenomena), like Roman or Carolingian scripts, or more ‘regional’, like the Insular or Merovingian, also came with elaborate visual-semiotic and linguistic systems. Simply put, abbreviations depended as much on scribal preferences as on the way the script looked (i.e. on the so-called aspect of the page) and the way it would have sounded (i.e. on the language of the page). Reference Ter Horst, Stam, Pahta, Skaffari and WrightTer Horst and Stam (2018) point to the ways in which Latin abbreviations were incorporated into eleventh- to fourteenth-century Irish manuscripts as visual diamorphs, that is, items whose graphic form prevents one from assigning them to a specific language (Reference Wright and TylerLaura Wright 2011: 203). For instance, the form aps. was devised to abbreviate Latin apostolus, but when applied to Gaelic, it would have rendered apstal (Reference Ter Horst, Stam, Pahta, Skaffari and WrightTer Horst and Stam 2018: 234).

Scribes copying in European vernaculars indeed adopted and adapted Latin bases for abbreviation to suit language-specific purposes. While on the whole, the orthographic ‘meaning’ of abbreviation symbols originating in Latin remained relatively unchanged throughout the Middle Ages, the morpho-lexical and visual contexts for those symbols provided a dynamic background for scribal abbreviating practices. The shift from Latin to vernacular languages necessitated the loss of some abbreviations, while others were successfully carried over, even if to encode purely orthographic, rather than grammatical categories, like they used to. As an example mentioned above, the -rum abbreviation for genitive plural of Latin third and fourth declensions in the form of figura ꝝ would no longer feature in vernacular scripts because it lacked a non-Latin equivalent in vernacular languages. On the other hand, an abbreviation like ꝭ – is, which marked the genitive plural of the same third declension in Latin, worked equally well as an indicator of plurality in English. Conversely, because some recurrent orthographic sequences in ‘national’ [sic ] scripts did not have their equivalents in Latin, scribes would devise new abbreviation symbols. Mixed-language manuscripts demonstrate how differently scribes were thinking about abbreviations in Latin and in vernacular languages – not only are the symbols distributed differently in linguistic terms (with Latin portions of the text typically containing more abbreviations than the vernacular fragments) but they are also visually different. Simply put: if the sheer number of abbreviations in the Latin text does not separate it from the vernacular fragments, then manipulating the type, size and/or color of the script surely does. In other words, it is through a combination of visual means (like the color of the ink, the size of the script, the type of the script, position on the page) and an increased incidence of abbreviations that switches between two (or more) linguistic codes are signaled (in the late fourteenth and fifteenth centuries, copies of works by John Gower or Geoffrey Chaucer are illustrative of these practices). The distribution of abbreviations on the page was also determined by their immediate word and line context. First of all, the position relative to other letters in the abbreviated word was significant, as that determined both the size and shape of the abbreviated symbol (whether or not it was attached to the preceding letter; if so, what letter that was, how much space there was between the x-height and the headline). Equally important was its position within the line: line-final and column-final contexts tend to attract abbreviations, whereas proximity to the right-hand margin or to the gutter of the page affects the dimensions of the abbreviation symbol. Marginal abbreviations (both in Latin and vernacular manuscripts) tend to occur with much more density and variety of types than those occurring within the ruled column, at the center of the page.

A more elaborate, formal type of script, like the textura, would encourage fewer abbreviations, as it would be typically applied to more costly manuscripts, often Latin ones, where saving time and space was not a concern.Footnote ⁵ In such scripts, abbreviations tend to be more formulaic in nature: they appear on items commonly abbreviated (like function words: prepositions, pronouns, conjunctions) or abbreviate words at the very ends of the ruled column so as to ensure visual coherence to the text on the page.Footnote ⁶ Vertical orientation of the script required the scribes to maintain larger distances between the lines (Reference 676DerolezDerolez 2003: 125). As a result, there was more interlinear space to fit abbreviation symbols, but because the script would be compressed, it was easier to fit words into the ruled column(s). Cursive scripts, in turn, are characterized by a more ‘open’ character and a horizontal orientation, which contributes to a more ‘sprawled’ aspect, giving scribes more trouble with keeping the text within the ruled grid and thus facilitating a more widespread use of abbreviations. Initially the development of the cursive did, indeed, contribute to a rapid expansion of abbreviation symbols, but when cursiva was adopted as a bookhand (rather than as a documentary script) in the fourteenth and fifteenth centuries, the number of abbreviations used in cursive manuscripts decreased significantly to facilitate the readability of the text and to enable access to the nonscholarly reader (Reference 676DerolezDerolez 2003: 154).

10.6 Abbreviations in Transition: From Manuscript to Print

Although abbreviations are commonly associated with manuscript books, their popularity and use continued well into the late fifteenth and sixteenth centuries, when the printing press was gradually becoming central to the book market. As pointed out by Reference Honkapohja, Liira and WrightHonkapohja and Liira (2020: 269), “[a]s parchment began to be replaced by a cheaper material, paper, and the printing press made it possible to produce multiple copies with ease, the two main needs for using an abbreviation and suspension system lost their importance.” Even so, abbreviations did not prove to be among the first casualties of “the printing revolution in early modern Europe” (see Reference EisensteinEisenstein 1983). Initially, printers attempted to present their codices in a format as visually similar to the handwritten book as possible: the mise-en-page (or, page layout), the illuminated initials, the illustrations, the incipits and explicits and the blackletter typeface, modeled on textura scripts, were all meant to look like typical elements of a manuscript. Matters were no different with abbreviations: such a familiar component of the codex would not go away with the arrival of new technology, and printers would sometimes go to extraordinary lengths to reproduce abbreviation symbols familiar from scribal practices (Reference Edwards and PearsallEdwards 2000: 65).

Before long, however, this labor-intensive process of book production proved untenable, as any advantage the printing press had over the man-hours devoted to handwriting was dwarfed by the costs (financial and otherwise) of printers’ efforts to replicate in metal casts the ligatures, variant graph forms and abbreviations (Reference Hellinga, Hellinga and TrappHellinga 1999: 70). Hellinga also points to the short-lived nature of founts of types which comprised a number of graphic variants for individual letter forms. These were gradually replaced by “long-lived type, carefully designed but with a smaller number of sorts,” which meant that in due course “a simplification in the presentation of graphic forms had to be accepted” (Reference Hellinga, Hellinga and TrappHellinga 1999: 71). In the end, abbreviations remained in use as one of three methods for right-justifying the text (the other two being breaking words over lines and altering spaces between words; see Reference ShuteShute 2017: 15), and several studies of the English material (e.g. Reference ShuteShute 2017, Reference Honkapohja, Liira and WrightHonkapohja and Liira 2020) have indeed pointed to a higher incidence of abbreviation in line-final contexts.Footnote ⁷ Similar motivation was noted in French texts (Reference CottereauCottereau 2005, Reference CampsCamps 2016, as quoted by Reference Honkapohja, Liira and WrightHonkapohja and Liira 2020: 298). Reference Honkapohja, Liira and WrightHonkapohja and Liira (2020: 301) link this development to a change in the layout of printed works from one to two columns of text, which exerted even more pressure on compositors to ensure right-margin justification (this change was also related to genre, as it was the de luxe books which were more typically printed in a two-column layout).

While mass publication increasingly became the norm for book production, various types of ephemera and legal or administrative documents continued to be handwritten (see Reference 713Kytö, Grund and WalkerKytö et al. 2011). For instance, in English, letters, personal notes, accounts, legal depositions and local documents (Reference Stenroos and 766WrightStenroos 2020a: 39 incorporates into this category “records of cities, churches, manors, local courts and private transactions”) remained the loci for abbreviation, oftentimes more elaborate than what was typical of handwritten manuscript copies (especially literary ones) toward the close of the fifteenth century. It is not without significance that documentary texts (“formulaic texts, and texts intended for internal use by administrators,” Reference Stenroos and 766WrightStenroos 2020a: 67) continued to be written in Latin well into the sixteenth century.

Accounts and inventories, on the other hand, often used abbreviations to obscure linguistic boundaries: this is what Reference Wright and TylerLaura Wright (2011: 203) referred to as the “visual diamorph” effect, by which abbreviation symbols applied in word-final (i.e. inflectional) contexts would often make it impossible to tell the difference between an English or a French or a Latin item. This visual overlap between Latin and the two vernacular languages used in medieval written accounts (Reference WrightWright 2002: 472) was one of the staples of code-intermediate phenomena, featuring in mixed-language texts prior to the emergence of a vernacular standard in the early modern period (Reference WrightWright 2002: 471). Yet another type of document characterized by a similarly prolonged use of abbreviations are the custom rolls. Like other types of handwritten accounts in late medieval England, these were heavily mixed-language texts with a system of abbreviations, which Reference Needham, Hellinga and TrappNeedham (1999: 152) referred to as “makeshift.”Footnote ⁸

Ultimately, “sometime in the sixteenth century” (Reference Honkapohja, Meurman-Solin and TyrkköHonkapohja 2013) abbreviations went out of use altogether. Reference Honkapohja, Liira and WrightHonkapohja and Liira (2020: 269) connect this development in England with the vernacularization of written documents. During the short-lived Commonwealth period of 1649–60, Latin abbreviations were abolished; they returned to use with the Restoration, only to be abolished again, along with Latin, in 1731 (Reference HectorHector 1966 [1958]: 29). Indeed, with the transition from monolingual Latin or Anglo-French through mixed-language to monolingual English in handwritten, and later, printed documents, only a handful of abbreviation symbols remained in active use and some of those are still used today (like e.g., i.e., etc.) as conventionalized markers of (often specialist) discourse.

10.7 Conclusion

Despite the fact that the modern reader would find it hard to infer this much from contemporary editions of historical texts (which tend to expand suspensions and contractions), abbreviations are intrinsically linked to pre-print literacy. Spanning late Antiquity through the early-to-late Middle Ages only to be laid to rest in the latter part of the early modern period, they belong to the commonplace community practices of premodern literati. Developed for the very pragmatic purpose of signaling oft-repeated words in legal and administrative records and of avoiding spelling out the nomina sacra, abbreviations became recognizable elements of scribal shorthand, which not only helped the scribes economize on their time and writing surface but were also employed in structuring the visual discourse (e.g. abbreviations were often used in running titles or in incipit/explicit sections to signal elements of text structure), signaling code-switching in multilingual texts (along with accompanying changes in ink color, type of script or position on the page) or, very generally, underscoring the Latinitas of the text (e.g. by reference to external authorities in the margins of the manuscripts). In a way, abbreviations are a truly protean element, not only changing in function but also in form: scribes adjust their shapes according to the amount of space available in the interlinear context and to the type of script selected for the copying process. In many ways they are a shorthand for medieval literacy: marrying the linguistic and the pragmatic, the visual and the textual, Latin and the vernacular through their multilayeredness and multiaspectuality.

11.1 Introduction

This chapter focuses on adapting different alphabetic writing systems to the requirements of a given language.Footnote ¹ For the needs of the present chapter, my main linguistic basis for analyzing writing systems are Slavic languages, with particular consideration given to the Polish language, which is used for detailed exemplification of the adaptation of writing systems under discussion. This chapter provides an analysis of examples of the way a variety of alphabetic systems were adapted to the Polish language, in order to showcase the complications behind the process of ‘recording’ (in the sense of rendering) a given language by means of a particular system of signs. The classical alphabet established for the Polish language consists in the appropriately modified Latin alphabet; however, one needs to realize the length of time involved in adaptation processes before a fully functional national alphabet takes shape. While users of a given language are typically unaware of this, any attempt at using an alphabet other than the classical alphabet makes clear the complexities of the mechanisms accompanying the alphabetization of a language. This process is presented here on the basis of several examples displaying the use of a different alphabet than the Latin alphabet for recording Polish: graždanka, the Arabic alphabet and the Armenian alphabet.

Adapting Lotman’s terminology, Reference OngOng (2005 [1982]: 8) deemed writing a “secondary modeling system” which is always dependent on the “prior primary system,” that is, the spoken language. Analogically, regarding the development of writing systems and indicating the ways in which particular alphabets have formed, one may introduce the term ‘prior modeling system’ referring to the alphabetic system on the basis of which the Ongian ‘secondary modeling system’ was formed – understood here as writing adapted to the needs of a given community. One may thus note two aspects of the way writing systems form within a certain language for a specific group of users: on the one hand, there is a transition from the oral system to the graphic system (prior primary system → secondary modeling system). On the other hand, the emergence of writing in a particular language concerns the very system of forming a set of signs, which never happens in isolation, as it draws on other, existing graphic systems, used by a different community, which usually communicates in a different language. As such, I consider the ‘prior modeling system’ a broad notion which, rather than a single system of signs, might contain several such systems; these systems, due to the influence of interferences, have given rise to a different alphabetic system. This aspect, in contrast to Ong’s cultural and social, or even philosophical reflections, emphasizes the technical issues pertaining to the graphic shape of linear alphabets.

The initial stages of the formation of almost every alphabet are characterized by a certain level of hybridity, resulting from adopting a borrowed graphic base and attempting to adapt it to one’s own needs. This stage tends to be marked by a diversity of technical solutions used to adapt the adopted graphic system that are imposed by features of the recipient language. As a result of further transformations and unifying actions, some solutions are forgotten, while, in time, others become established methods of alphabetic writing in a given language (national alphabet).Footnote ² In this chapter, the broadly understood hybrid nature of writing systems pertains to every situation wherein an alphabet originally serving to record statements in a different language was adopted at an early stage of its adaptation connected with introducing new and individual graphic solutions. Conversely, writing systems that were genetically connected with a specific language, or a language family, and then entered a stage of relative stabilization are considered base systems; however, at an initial stage of their formation they might also have developed on a different base graphic system.Footnote ³ Fully formed systems are most frequently characterized by centuries-long traditions of using a specific system of signs for a given language, and its gradual evolution. The most desirable result of such outlined changes is the best possible regulation of the phone–sign relationship.Footnote ⁴

Below, in Section 11.2, I present the process of the crystallization of alphabetic systems, and the significance of contacts and interferences between them. In Section 11.3, the hybridity of both writing systems and orthographies is considered on the level of the formation of a national writing system, while in Section 11.4, it is analyzed from an even narrower perspective. I discuss here particular cases of transposition and utilization of a different alphabet from that used as a standard in a given language.

11.2 Alphabet Formation: Adaptations and Interferences

Without fully investigating the genesis and the stages of the formation of an alphabetic system of recording speech, it may be assumed that they began to appear during the second millennium BC in Western Semitic writing (see, e.g., Reference DiringerDiringer 1943, Reference StrelcynStrelcyn 1952, Reference CohenCohen 1956: 45–46, Reference GelbGelb 1969: 190–205, Reference 689GieysztorGieysztor 2009: 40), although it is very likely that graphic elements of earlier pictographic origin were used and transposed (Reference CohenCohen 1956: 44, Reference 689GieysztorGieysztor 2009: 41). Indisputably, one of the most significant achievements in the development of writing is the creation of Phoenician writing, with its abjad alphabetFootnote ⁵ and its considerable influence on the development of other graphic systems.Footnote ⁶ This section discusses two examples that document the adaptation and changes in the base systems of signs.

The assumptions outlined in Section 11.1 may be exemplified by the development and evolution of the Greek alphabet, which was based on the Phoenician alphabet. In order to record their language, the Greeks modified the system previously used by the Phoenicians (Reference CohenCohen 1956: 50). Although particular city-states used local varieties of the adapted letters, a common feature was that the Phoenician consonantal alphabet took on the form of a full consonant-and-vowel alphabet in its Greek variety. To mark vowels, Phoenician letters denoting phonemes absent from the Greek language tended to be used, and several signs unknown to the Semitic prototype also appeared. Neither the Phoenician nor the early Greek writing separated words; it was only with time that Greek writing abandoned this continuous style of writing (Reference CohenCohen 1956: 45, 52–53). In contrast to the Phoenician script, which had a fixed direction from right to left, only some of the oldest records retain this direction, while others are written boustrophedonically (Reference CohenCohen 1956: 45, 51), a manner, discarded in approximately the sixth century BC, for a single-direction arrangement from left to right. The present-day Greek alphabet is a bicameral script with a clear differentiation between uppercase and lowercase letters.Footnote ⁷

An even more interesting example may be the emergence of the alphabet known as the Glagolitic alphabet (from glagolъ ‘word, letter’ in Old Church Slavonic, OCS for short; see Reference BrajerskiBrajerski 1990: 58) and of its subsequent transformations. The creation of this alphabet is usually attributed to the missionary brothers Cyril and Methodius, who arrived in the vicinity of Great Moravia with a mission to Christianize the region in the second half of ninth century (see also Chapter 8, this volume). They recorded the translated liturgical texts in an alphabet created to reflect the phones of the OCS, the contemporary Slavic dialect of Solun (a Bulgarian-Macedonian dialect from the region of Thessaloniki). The prototype of the letters of the Glagolitic alphabet is thought to be found, above all, in the Greek minuscule of the eighth and ninth centuries, although features of Southern Semitic alphabets and Latin cursive are also frequently pointed out.Footnote ⁸ The letters were written, as in contemporary Greek, horizontally from left to right. This graphically uniformly stylized alphabet should, however, certainly be considered an original creation, individually tailored to the phonetic and phonological needs of the language it was intended to record.Footnote ⁹ The original Glagolitic alphabet underwent certain modifications concerning the shape of its letters; for example, in Bulgaria the so-called round Glagolitic alphabet persisted, while in the Croatian Primorje region and Istria the so-called angular Glagolitic alphabet developed, most likely under the influence of the Gothic script. In the tenth century, a new alphabet began to spread in the place of the Glagolitic alphabet, the Cyrillic alphabet (so called to honor Cyril), whose basis is believed to be found in the majuscule of the Greek alphabet (in its uncial version)Footnote ¹⁰ together with simplified Glagolitic letters denoting phones without equivalents in the Greek alphabet (Reference DiringerDiringer 1948: 478–83, Reference Lehr-Spławiński and BartulaLehr-Spławiński and Bartula 1976: 7). Although genetically both scripts under discussion show specific dependencies, the Glagolitic alphabet with the minuscule and the Cyrillic alphabet with the majuscule, both are bicameral scripts, with the uppercase and lowercase letters differing not in shape, but in size.

11.3 National Scripts and Alphabets in a Variety of Linguistic Editions

The contacts and interferences between alphabets take place on two planes. The basic plane, as discussed in Section 11.2, showcases the stages of the formation of alphabets such as the Greek, Latin, Cyrillic and Arabic ones, taking place with the use, adaptation or transformation of basic systems of writing. These may be described as the ‘grand’ alphabets: not only the ones most commonly used but also constituting the basis for a number of variants (i.e. national scripts). On this plane, we are dealing with already specific adaptations of an established alphabet for the needs of a given language, with the use of conventionalized methods of modification. The formation of national scripts, that is, individualized systems of alphabetic writing adjusted to the needs of given peoples, usually consists in variations of adaptations of the most frequently used alphabets. One such basic European alphabet is, above all, the Latin alphabet, in its original form used to record the Latin language. The modifications of this alphabet for the needs of other languages, such as Romance, Germanic, Finno-Ugric and Slavic, may be schematically presented as follows:

AL—L → ALX—Xwhere the Latin alphabet (AL) used in the Latin language (L) is adapted for recording another language (X) with the use of the adapted Latin alphabet (ALX).

The appropriate adaptation of the adopted basic alphabet (AL → ALX), suitable for recording language L, to the particular needs of language X would take place in several ways. To graphically represent a phone from language X that is absent from the phonetic system of language L and, thus, has no graphic equivalent therein, required several resolutions. The basic elements of new graphemes were, above all, graphically or functionally modified letters from the original alphabet: (a) letters with diacritics (usually representing a phonic value close to the original letters), for example with signs used originally to mark different accents: dots, curves, strokes, curls or double marks; (b) letter combinations (digraphs, trigraphs, etc.) or ligatures corresponding to a single phone of language X; (c) letters whose shape has been modified; (d) letters that have no phonic equivalents in language X in a different function from that of language L. A less common way of acquiring graphemes reflecting the specificity of the recipient language (X) consists in borrowing letters from alphabets different from AL, whereas entirely new letters were created only sporadically (see also Reference DiringerDiringer 1948: 553–54). The methods by which the basic script would be adapted to the needs of the national scripts presented here pertained not only to the Latin alphabet, or even to national scripts as such, but also other ones (see the adaptations of the Latin, Cyrillic, Arabic and Armenian alphabets discussed in Subsections 11.3.1, 11.4.1, 11.4.2 and 11.4.3).

In our discussion of national scripts we may use the analogy to OCS, which became the language of the liturgical writing system for Slavs from the Byzantine culture. At the same time, it absorbed elements of a number of local dialects. We are therefore dealing with the same language, but in different permutations, in academic literature known as editions, for example OCS in the Ruthenian, Croatian and Bulgarian editions. A similar situation pertains to alphabets. In a given culture’s sphere of influence, the adapted alphabet would function to record other national languages, and would later be modified to the needs of their own writing system. Analogically to languages, we may thus refer to alphabets in a given national edition, such as the Latin alphabet in a German, Czech, French or English edition, or the Cyrillic alphabet in a Belarusian, Macedonian or Bulgarian edition. The Cyrillic alphabet, presented in Section 11.2, underwent subsequent stages of evolution. Peter the Great’s reform at the beginning of the eighteenth century introduced a modified and graphically modernized version of the Cyrillic alphabet to Russia, the so-called graždanka, or civil script (from a Russian adjective in the term гражданский шрифт ‘civil, secular script’). Over the following two centuries the inventory of the letters in the script changed. Some letters that had no equivalent in Russian were omitted (e.g. letters used to denote nasal vowels in OCS) and new letters were added to better differentiate between phones (e.g. to denote iota).

With certain modifications modeled on the Russian alphabet, the Cyrillic alphabet became the basis for the alphabet that is used today in, among other countries, Bulgaria, Serbia, Montenegro, Macedonia, Bosnia, Herzegovina, Moldavia, Belarus and Ukraine. After 1991, some former republics of the USSR (e.g. Turkmenistan, Azerbaijan, Uzbekistan)Footnote ¹¹ ceased to use the general Cyrillic alphabet and adapted the Latin alphabet to the phonic needs of their languages. In Bosnia and Herzegovina, as well as in Uzbekistan, both the Cyrillic alphabet and a modified Latin alphabet are used, while in Tajikistan the Persian alphabet is also used. The oldest alphabet in Croatia, the Glagolitic alphabet, was long used, but for a considerable period of time Croatian writing was characterized by its tri-alphabeticity, as simultaneous use was made of the Glagolitic, Cyrillic and Latin alphabets. In Czechia, the Glagolitic alphabet was superseded by the Latin alphabet as early as at the end of the eleventh century (Reference OczkowaOczkowa 2004: 57). This brief overview of the evolution of alphabets, mainly those used to write Slavic languages, simultaneously showcases fluid changes from the original graphic systems to systems of a hybrid nature, which, subsequently, having gone through a number of stages of adjustment, begin to constitute a system congruent with the phonetic layer of the given language. Against the background of these examples, the Glagolitic alphabet stands out as the alphabet whose functional perfection is already confirmed in the oldest inscriptions.

11.3.1 The Polish Edition of the Latin Alphabet

To record the Polish language, the Latin alphabet began to be used in the thirteenth and fourteenth centuries. The letters available, however, were insufficient to represent all the phones in the Polish language, and thus only when the Latin alphabet was considerably expanded and a set of orthographic rules added was it possible to record Polish phonically adequately and unambiguously. The development and expansion of printing in the sixteenth century played a significant role in the standardization of the rules of Polish writing, although the creation of a unified and consistent system was still a long way off.Footnote ¹² An analysis of the way vowels were written at the time adequately demonstrates the methods of adjusting the Latin alphabet by dint of the adaptative techniques discussed above (see Section 11.3).

A more extensive vocal system required the introduction of new graphemes. Most frequently, this consisted in using a suitable letter with a diacritic. For nasal vowels denoting a nasal version of the /ɔ/ and /ɛ/ vowels, the letters a and e were written with a diagonal stroke at the bottom of the stem, or, less frequently, with an ogonek (littera caudata);Footnote ¹³ indeed, the latter method persists in the modern Polish alphabet: <ą>, <ę>. Earlier and for a long time, the Polish language had only one nasal vowel of a quality approximate to that of the /a/ vowel. In order to denote this, Reference Parkosz and KucałaParkosz (c. 1470: 44, Kraków, Jagiellonian Library, MS 1961, pp. 3–16),Footnote ¹⁴ author of a medieval orthographic treatise, recommended the use of <ø>, which was already available in this function at the end of thirteenth century. In the sixteenth century, former quantitative differences (duration) between vowels were replaced with a differentiation in sound. Former long vowels, usually denoted by double letters (e.g. long a as <aa>; see Reference PallasParkosz c. 1470: 44) in medieval texts, underwent raising in the Renaissance (the sixteenth century), becoming what are now called leaning or narrow vowels. The quality of short vowels (so-called bright vowels) did not change in pronunciation.Footnote ¹⁵ Two such vowels, /ɛ/ and /ɔ/, continued to be written with unchanged graphemes, while the /a/ vowel was usually written with an acute accent <á>. New vowels with raised articulation were written, respectively, as <é>, <ó> and <a>Footnote ¹⁶ (see Reference JanuszowskiJanuszowski 1594: E). This notational system may be considered the most typical, but other methods were also utilized. In a Renaissance description of the Polish alphabet in three different lessons proposed by three humanities scholars of the time, Jan Kochanowski, Łukasz Górnicki and Jan Januszowski (Reference JanuszowskiJanuszowski 1594),Footnote ¹⁷ we also find different suggestions for diacritic signs with letters denoting vowels, for example in Kochanowski’s version /a/ : <á>, /ä/ : <à>;Footnote ¹⁸ /ɛ/ : <é>, /ɛ̇/ : <è>; /ɔ/ : <o>, /ɔ̇/ : <ò> or <ó> (Reference JanuszowskiJanuszowski 1594: G3). In the earlier texts of this period there were sometimes instances of a notation parallel to that discussed above, denoting a narrow vowel by means of doubling the letter, and sometimes additionally with a diacritic sign, for example /ä/ : <aa>, /ɛ̇/ : <eé>, /ɔ̇/ : <uó>. Another sporadically utilized method of adapting the Latin alphabet to denote vowels of the Renaissance Polish language was borrowing from another alphabet. The grapheme <α> (derived from a Greek letter), for instance, could be used as an equivalent of the Polish nasal front vowel, while the grapheme corresponding to the same letter with a stroke across the stem, <α_̷>, represented the nasal back vowel. In present-day Polish, there remain two nasal vowels /ɔ̃/, /ɛ̃/, written by means of a letter with an ogonek <ą>, <ę>, and oral vowels in a single articulative variant, so there is no further need to differentiate <a>, <e>, <o> (or <i>, <y>, <u>) graphically. The two formerly narrow vowels are currently pronounced in the same way as their bright equivalents /a/, /ɛ/, while the previous narrow o /ɔ̇/ is pronounced as /u/. Historical spelling has, however, been retained and the /u/ vowel derived from the old leaning /ɔ̇/ vowel is denoted by the grapheme <ó>.

Sixteenth-century texts also make use of other methods of adapting the Latin alphabet with respect to consonants: ligatures, modifications of the letter shape, changes in the function of a letter, and the creation of new letters (together with the creation of new ligatures). The most common ligature consisted in the grapheme borrowed from German script (with a change in function), resembling the Greek letter β, the so-called Eszett or scharfes S (sharp s). In the Polish writing system, it denoted a particular front-tongue postalveolar consonant /ʃ/,Footnote ¹⁹ written in two variants depending on the typeface: one close to the German original <ß> (e.g. Reference JanuszowskiJanuszowski 1594: G3, H3 v) and, more frequently, one consisting in a ligature combination between a long s and a z with a cauda <ſʒ>. Górnicki (Reference JanuszowskiJanuszowski 1594: G3 v) proposed his own ligature for the consonant /ʧ/, consisting of two connected c letters. Like Kochanowski, he also advocated the use of a ligature for the voiced equivalent of this consonant, modeled stylistically on German ligatures (Reference JanuszowskiJanuszowski 1594: G3–G3 v). Kochanowski denoted the soft /ɕ/ consonant either with the letter s with an acute <ś> or with a considerably modified letter s with a closed lower bowl, approximating in its shape the Arabic digit 6 written in cursive (Reference JanuszowskiJanuszowski 1594: G3). In Reference CzerneckiCzernecki’s (1902: 12) opinion this was a new grapheme created by Kochanowski. As suggested by Reference Parkosz and KucałaParkosz (c. 1470: 45–47)Footnote ²⁰ as early as during the Middle Ages, differentiating between hard and soft consonants on the basis of differentiating the shape of the letters generally failed to catch on. It found its continuation, however, in the sixteenth-century Ortografia Polska by Reference MurzynowskiMurzynowski (1551). Like Parkosz, this author also suggested the use of the letter b with a top bow to denote the palatal consonant, but only in word-final position, while he denoted the palatal /ʑ/ phoneme by means of <ɀ> (z with a ‘swash tail’) (Reference MurzynowskiMurzynowski 1551: Bv, B3).

Figures 11.1, 11.2 and 11.3 show three Renaissance suggestions for a Polish national alphabet (including allographs in Januszowski’s version):

Figure 11.1 Alphabet according to Jan Kochanowski

(Reference JanuszowskiJanuszowski 1594: G3, CBN POLONA online library, https://polona.pl)

Figure 11.2 Alphabet according to Łukasz Górnicki

(Reference JanuszowskiJanuszowski 1594: G3 v, CBN POLONA online library, https://polona.pl)

Figure 11.3 Alphabet according to Jan Januszowski

(Reference JanuszowskiJanuszowski 1594: H3 v, CBN POLONA online library, https://polona.pl)

The current Polish alphabet consists of 32 letters:Footnote ²¹ A a, Ą ą, B b, C c, Ć ć, D d, E e, Ę ę, F f, G g, H h, I i, Jj, K k, L l, Ł ł, M m, N n, Ń ń, O o, Ó ó, P p, R r, S s, Ś ś, T t, U u, W w, Y y, Z z, Ź ź, Ż ż, of which nine have diacritic signs.Footnote ²² Every letter corresponds to a specific phone with distinctive features. The graphemes <u> and <ó>, however, represent the same phoneme, /u/, and their usage is motivated by morphological and historical reasons. Digraphs <ch>, <cz>, <dz>, <dź>, <dż>, <rz>, <sz> which, with the exception of particular cases, denote the following seven phones: /x/, /ʧ/, /ʣ/, /ʥ/, /ʤ/, /ʒ/, /ʃ/, are not included in the Polish alphabet.Footnote ²³ Soft consonants, denoted by a letter with an acute accent, are written without a diacritic if they occur before a vowel, and their palatal character is denoted by a following, additional letter i,Footnote ²⁴ with which they create two- or three-piece signs: |ć| ~ |ci|, |ń| ~ |ni|, |ś| ~ |si|, |ź| ~ |zi|; |dź| ~ |dzi|. In contrast to the aforementioned digraphs, they correspond not to separate sounds, but constitute only orthographic (positional) allographs of appropriate diacriticized letters.

11.4 Alphabet Adaptations with a Narrow Scope of Usage

The above-discussed adaptations of alphabets aimed at rendering a particular language may be considered to represent an adaptation of a general range. They are, firstly, connected with a top-down regulation regarding the usage of a specific alphabetic system, and subsequently, with refining the relationship between the phonemes and the graphemes, and with formulating the orthographic rules. This is the way national alphabets and scripts may be born. Due to their national nature, they have a broad group of users. In many countries, the rules of writing are regulated officially (see the example from Footnote note 22); they are relatively highly normalized, and they may frequently be characterized by gradual evolution, resulting from long tradition. Writing “as a public good” (Reference CoulmasCoulmas 2013: 104) is a source in the public sphere in many aspects, including the institutional. There also exist examples of adaptations of alphabets whose range of usage is considerably narrower than those described in Section 11.3 above. In such cases, we are dealing with a different kind of hybridity of the writing systems than that connected with the early stages of adjusting an alphabet to the general needs of a community that speaks a particular language. This pertains to scripts that only partly fulfill the same criteria as do national scripts. In some republics of the former Union of Soviet Socialist Republics, the lack of a longer writing tradition stems, above all, from political conditions. Due to the relatively frequent changes in the general alphabetic system of the languages that prevail there, as well as the ordinances governing the official language, many of the alphabets imposed have failed to proceed from the stage of early adaptation to the level of solid stabilization, as had been the case in the aforementioned systems of writing for the Turkmen, Azeri and Uzbek languages (see references in Footnote note 11).

There may be a variety of reasons behind adaptations of a narrow scope of usage, that is, adaptations of a provisional rather than general nature, for example, political, ideological, religious adaptations. There may also be different dependencies between alphabets used in a general scope and in a narrow scope. On the one hand, an alphabet utilized provisionally in order to record a given language may already be broadly used to record other languages in the same language group, as in the case of the traditional use of the Cyrillic or the Latin alphabet to record Slavic languages. On the other hand, it is also possible to transpose an alphabet that generally serves to record languages from an entirely different language group (see Subsection 11.4.2 about aljamiado). The difficulties connected with transposition of this kind should be considered on two levels. The first pertains to the differences of an interlanguage nature, among which the most significant are differences in the phonological subsystems of languages from distantly related language groups: a different stock of phones, different realizations of phones, their distinctive functions, positional variants, etc. The very systems of writing are frequently different (e.g. full alphabet or abjad). There are other significant differences of an external nature, such as the prevalent writing practices, orthographic rules and the direction of writing, but these do not receive any further attention here due to space limits.

11.5 Conclusion

This chapter has focused, above all, on the stages of the hybridity of writing systems during the formation of various ways of writing, that is, the adaptation of a certain alphabetic system (the base alphabet) for the needs of the phones of a different language. I have pointed out various paths of adaptation, different scopes and diverse reasons for adopting one and not another base system. In Section 11.2, attention was drawn to the crystallization of ‘grand’ alphabets (Latin, Greek, etc.). The discussion then moved to the modifications of established base systems, which sought to create a well thought-out, uniform system of rendering of another language. I focused on two levels of adapting base alphabets for the needs of other languages. Broadly, the adaptation of writing is of a universal nature, common to the languages discussed in Section 11.3, as was seen from the example of the Latin alphabet in its Polish edition (11.3.1). The adaptation of an alphabet of narrow usage scope (Section 11.4) for various reasons would not achieve the status of a national script. I analyzed the following examples: the Polish Cyrillic alphabet (11.4.1), the Polish and Belarusian aljamiado (11.4.2) and the use of the Armenian alphabet to write Polish texts (11.4.3). While the examples of national alphabets (Section 11.3) and alphabets of narrow scope of usage (Section 11.4) concern the recording of the Polish language, the universal nature of certain methods related to adaptation techniques of alphabetic systems also needed to be addressed. The aforementioned ways of recording the Polish language illustrate an entire spectrum of dependencies connected with choosing which alphabetic system should be used: cultural, religious, political, social or utilitarian.

12.1 Introduction

This chapter provides insights into crucial findings on variation and change in orthography and into current theoretical and methodological issues.Footnote ¹ It gives an overview of types and functions of historical orthographic variation in Section 12.2 and different patterns of orthographic change in Section 12.3. Its aim is to introduce the subject and to highlight important preliminaries, methodological considerations, relevant research questions and desiderata. This chapter shall, on the one hand, contribute to a broader general understanding of orthographic variation and change in historical texts and, on the other hand, provide a background against which a research question and design for the study of variation and change in historical orthographies can be defined and created.

The topics covered in this chapter are limited in three important respects. First, it includes only alphabet writing, that is, writing systems with distinctive segments that systematically map to phonic units.Footnote ² Second, the chapter concerns mainly the history of German and runic writing.Footnote ³ Lastly, within the history of German, the main focus rests on the Middle Ages (AD 750–1350) and the beginning of the early modern period, before convergence processes toward a supraregional variant began. These restrictions reflect the focus of my own research. However, the examples from two writing systems and different time periods provide insight into various processes and contexts for orthographic practice and I hope that the conclusions on historical orthographic variation and change presented here are transferable to other historical written languages. Moreover, insights into variation patterns and developments in premodern writing are not only relevant for linguistic sister fields, particularly historical phonology (see the chapters by Reference Lass, Honeybone and SalmonsLass 2015, Reference Minkova, Honeybone and SalmonsMinkova 2015a and Reference Unger, Honeybone and SalmonsUnger 2015 in the Oxford Handbook of Historical Phonology), but can also be of interest from an interdisciplinary perspective, given that historical texts are investigated in other disciplines, including paleography, history, literary studies and cryptography.

In my description of written language, I use the following technical terms: sign and letter to refer to the basic, visually distinct, individual units in an alphabetic writing system, such as the Roman alphabet letters a, b, c, d and the runic signs ᚠ, ᚢ, ᚦ, ᚬ; graph (German: Graphie) or graph type for signs or a combination of signs that map to a phonic unit of a reference variety;Footnote ⁴ and grapheme for the combination of a graph and the phonic unit of the historical reference system it represents in a particular spelling system.Footnote ⁵

12.2 Variation in Orthography

Spelling variation is one of the key characteristics of premodern writing. Various spellings of the same word are common not only in a language of a certain period, for example the spellings kind, chind, chint, khind, kint, kynt, kindt, kinth for the lemma Kind ‘child’ in Early New High German,Footnote ⁶ but can also occur in the same text written by the same scribe. While modern written language users are familiar with linguistic variation on the grammatical and the lexical levels, spelling variation is in sharp contrast with most present-day, standardized orthographies. Here, invariability in word spelling and a predefined set of letters and graphs are crucial principles.

In traditional linguistic research, spelling variation in historical texts was perceived and empirically approached through the lens of present-day orthography. It was often classified as chaotic and random and the scribes were judged as inconsistent and careless.Footnote ⁷ More recent research has shown, however, that historical orthographic variation is not only an integral part of historical languages, but also follows systematic patterns and has functional uses. Premodern orthographic systems are not standardized and, therefore, there is room for innovation. This is a definitive characteristic and an important precondition for orthographic variation and change. Hence, premodern systems and uses of written communication were very different from present-day orthography. This “otherness” (Reference MihmMihm 2016) has to be taken into consideration in approaching historical orthographies, and “our job is to develop a hermeneutic that non-anachronistically provides techniques for anything we are likely to come across in past writing” (Reference Lass, Honeybone and SalmonsLass 2015: 107).

While linguistic variability as an inherent characteristic of all languages is widely investigated on the levels of phonology, morphology, syntax and lexicon, the study of orthographic variability has attracted somewhat less attention. Interest in the structural description and functional explanation of orthographic variation patterns has grown only in the last decades. Consequently, theoretical and methodological aspects of the study of orthographic variation are still subject to ongoing development and refinement. In this section, theoretical and methodological challenges for the study of orthographic variation in historical documents as discussed in recent studies are presented. This discussion is followed by an overview of approaches to different types and functions of orthographic variation using examples from the history of alphabetic German and runic writing.

12.3 Orthographic Change

This section begins by providing an overview of general tendencies in orthographic change. Two case studies are then presented, the first from the history of alphabetic German and the second from runic writing, that illustrate different structural paths of orthographic development and their triggering factors.

12.3.4 Graphemic Change in Runic Writing

In the history of runic writing, two opposite trends can be observed: developments toward more phonological determination on the one hand and toward a more abstract, reductive system on the other hand. While the former developments have been linked to the external factor of a multilingual and multiscriptal, Latin-learned milieu, explanations for the latter trend are based on internal, structural conditions. A case for the shift to a phonologically more determined runic orthography is the development from the common-Germanic elder futhark to new versions in Anglo-Saxon England. On the basis of a sound-relational analysis of the Old English runic corpus, Reference Waxenberger, Waxenberger, Sauer and KazzaziWaxenberger (2017) divides the process of change into three phases. The triggers in the first phase (c. AD 400–600) were phonological changes in the vowel system (e.g. the phonemic split of West Germanic */aː/ into pre-OE /aː/ and /ɔ̃ː/) that led to changes in the graphemic system. New symbols were created by graphic modification (e.g. ᚩ for /ɔ̃(ː)/ and ᚪ for /a(ː)/), and former signs changed their sound values (e.g. ᚨ from */a(ː)/ > /æ(ː)/ and ᛟ */o(ː)/ > /ø(ː)/). While these changes are evident throughout the Old English runic corpus, the innovations of the second phase (c. AD 650–750) clearly originate in one inscription, the Ruthwell Cross in Northumbria, a wooden cross featuring a devotional Christian poem in Old English. The innovations include a new sign ᛠ for the diphthong /e(ː)a/, three completely new creations and a recycled use of an older sign ᛇ to render products of velarization and palatalization. Only the use of ᛠ for /e(ː)a/ spread to southern regions, whereas the other innovations were either restricted to the Ruthwell Cross or only influenced a few neighboring inscriptions in the North-West of Northumbria (see also Reference WaxenbergerWaxenberger forthcoming). The third and final phase is characterized by the recycling of older runic signs that had become superfluous to distinguish subphonemic palatal and velar consonant allophones in different sound positions. Reference Waxenberger, Waxenberger, Sauer and KazzaziWaxenberger (2017: 244) points out that the later innovations clearly “were carried out by learned scholars.”

The social embedding of the later Old English innovations in a Latin-based, learned, Christian context is emphasized by Reference SchulteSchulte (2015: 97) as a crucial external factor which is absent in the runic developments in the Scandinavian context around AD 600–700. In the North, the runic sign inventory was significantly shortened from 24 to 16 signs (see Figures 12.1 and 12.2) and, as a consequence, a number of graphs mapped to more than one sound value, that is, became polyfunctional (e.g. ᚢ ‘u’ – /u(ː)/, /y(ː)/, /o(ː)/, /ø(ː)/, /w/; ᛏ ‘t’ – /t(ː)/, /d(ː)/). The outcome clearly is an economical writing system. Additionally, in contrast to the Old English changes in the sign inventory, the innovation strategies in Scandinavia did not include either the creation of completely new or the recycling of superfluous signs, but rather loss of signs and graphic simplification. Reference SchulteSchulte (2011, Reference Schulte2015: 101–2) argues that the developments that led to the younger futhark in the North are part of an invisible, structural change based on the writing principle of polyfunctional runes and resulting in “the abandonment of phonemic distinctions and the creation of maximum contrast” (Reference SchulteSchulte 2015: 101). Reference SchulteSchulte (2011: 57–60) sketches the development as originally being triggered by a phonological change, the umlaut, which led to the polyfunctional use of the rune ᚢ ‘u’ to render both /u(ː)/ and /y(ː)/. This polyfunctionality of a single sign would have developed into a writing principle and was analogically transferred to the consonant system, diminishing the distinction between voiced and unvoiced plosives (ᛒ ‘b’ for both /p/ and /b/, ᛏ ‘t’ for /t/ and /d/, ᚴ ‘k’ for /k/ and /g/). The exact course of the change and plausible explanations for single developments involved are still lacking within runology. It is, for instance, unclear why certain signs were abandoned while others changed their sound value (Reference BarnesBarnes 2012: 54–59). Furthermore, there are no explanations for the diffusion of the changes throughout Scandinavia. A concise, sound-relational analysis of the ‘transitional’ inscriptions as well as the consideration of geographical and, as far as possible, chronological and other external factors might shed new light on variation patterns and the diffusion of innovative spellings. Such an analysis might also add to the discussion on whether the change occurred gradually as a natural language change or rather constitutes a planned reform ‘from above’, or a combination of both.

Figure 12.1 The elder futhark consisting of 24 signs (c. AD 100–700)

Figure 12.2 Swedish variant of the younger futhorc consisting of 16 signs (c. AD 700–1200)

In the 1000s AD, structural changes aiming at differentiating polyfunctional signs emerged in Scandinavian runic writing. This was mainly achieved by adding diacritic dots to render distinctions between, for example, ᚢ ‘u’ /u/ and ᚤ ‘y’ /y ~ ø/, ᛁ ‘i’ /i/ and ᚽ ‘e’ /e/, ᚴ ‘k’ /k/ and ᚵ ‘g’ /g/. The origins of this practice are still obscure even though an emergence in the multiscriptal (runes, Roman alphabet, ogham)Footnote ¹² and multilingual (Latin, Celtic, English, Scandinavian) context of the British Isles, where at that time there was lively exchange with Denmark and Norway, seems most appealing (Reference Hagland, Page, Dybdahl and HaglandHagland and Page 1998, Reference SchulteSchulte 2020). The practice spread to the whole of Scandinavia, but it is worth mentioning that it was not employed consistently; ᚴ could still denote /g/, for instance. Moreover, the dotted characters were integrated only very seldom in so-called ‘futhark inscriptions’, inscriptions where the runic signs were listed as an inventory in their traditional sequence. This fact suggests that the additional characters were considered to be spelling variants by the carvers (see Reference Spurkland, Waxenberger, Sauer and KazzaziSpurkland 2017). At the same time, further changes occurred that contributed to a more accurate rendering of phonological distinctions. The use of these spelling innovations is usually linked to the rise of Latin literacy and the introduction of the Roman alphabet in the North in the later Viking Age. However, a recent study of Swedish runic inscriptions in Latin (Reference Palumbo, Marold and ZimmermannPalumbo 2022) has shown that the “encounter between the Latin and the runic written culture […] is a multifaceted phenomenon where processes are at work that are more intricate than a simple one-way influence from one tradition to another.” Among other sociolinguistic factors, spelling practices rely on different levels of competences among carvers.

12.4 Future Prospects

Two recent systematically executed sound-relational graphemic studies on runes, both of which follow a corpus-oriented approach and take a macro perspective (Reference Waxenberger, Waxenberger, Sauer and KazzaziWaxenberger 2017, Reference PalumboPalumbo 2020, see Subsections 12.2.5 and 12.3.4), showed chronological and geographical patterns in spelling variation and identified crucial phonological and cultural factors behind spelling variation and change. Systematic research on orthographic variation and change is, however, still rather sparse. For German, research is moreover confined to micro-studies covering single scribes or smaller groups of scribes and typesetters in certain regions and in certain time periods. On this basis, it is not yet possible to give a more detailed and general picture of spelling variation patterns and diachronic developments of historical orthography, and the explanation for some patterns occurring in micro-studies must remain open due to the lack of comparable data. There is a need for both more micro-level studies for regions and contexts not yet covered in previous research and comparative studies covering larger datasets. The methodological approach in the Atlas of Late Medieval Written Languages of the Low German ‘Altland’ and Neighboring Areas (Reference PetersPeters 2017) might serve as a model for projects covering other regions, or even larger areas. It is based on dated and geographically located original sources, and both synchronic variation and diachronic developments are visualized and thoroughly documented. Furthermore, such systematically collected and presented data on tendencies in orthographic variation and change can serve to date and locate manuscripts (see, e.g., Reference Rutkowska and RussiRutkowska 2016: 186, Reference SeelbachSeelbach 2016).

Challenges for future research certainly lie in the complexity of the interplay between different factors for spelling variation and the various functions single spellings may perform. This applies especially to large-scale projects that become laborious and time-consuming when factors such as origin, provenance, dating, the scribe or typesetter and their educational and professional background, the possible presence of an exemplar, and other philological factors must be collected and evaluated for every text witness. Moreover, metadata are often not available, especially in epigraphic data such as runic inscriptions, where a scribal separation only can be carried out in rare cases. However, future studies benefit significantly from the availability of a growing number of reliable editions that are also accessible digitally (see, e.g., the collection of historical corpora in the CLARIN ERIC infrastructure and the HistCorp collection presented in Reference Pettersson, Megyesi, Mäkelä, Tolonen and TuominenPettersson and Megyesi 2018). Such digital corpora might, in the future, even be applied to reduce the number of mistakes in optical character recognition (OCR) and in handwritten text recognition (HTR)Footnote ¹³ of historical texts where spelling variation still poses problems.

Another way forward in the analysis of orthographic variation and change are digital methods. An example where digital methods are exploited to good effect is Reference KorkiakangasKorkiakangas’s (2018) study on spelling variation in administrative Latin writing from seventh- to eighth-century Tuscany. Korkiakangas applies the Levenshtein distance (see Reference LevenshteinLevenshtein 1966) whereby the deviation of an actual word spelling from a normalized type is measured automatically (for a thorough discussion of the approach, including its limitations for historical material, see Reference KorkiakangasKorkiakangas 2018: 577–81). The computer-assisted correlation analysis is visualized in a series of different diagrams (see also Reference Korkiakangas and LassilaKorkiakangas and Lassila 2018) and has shown both significant interwriter variation throughout the material and a reduction in variation over time. Lastly, the field of historical orthography would profit from more comparative studies on variation and change not only in different vernacular languages (see Reference Voeste, Baddeley and VoesteBaddeley and Voeste 2012b, Reference SeilerSeiler 2014, Reference NowakNowak 2019, Reference Dücker, Hartmann, Szczepaniak and CondorelliDücker et al. 2020: 86–89) but also taking practices employed for contemporary Latin writing into consideration.

12.5 Conclusion

This chapter has given an overview of current tendencies in the research of variation and change in historical orthography and has outlined relevant methodological considerations for its study. Examples from historical varieties of alphabetic German and runic writing have served to illustrate different types and functions of spelling variation and tendencies in graphemic change. In current research, micro-studies prevail and, to date, only little can be said about larger processes and about similarities and differences of orthographic variation and change in different languages. More research is needed in order to shed light on larger tendencies in variation patterns, on their functions and on the historical development of orthographic practices. Moreover, comparative perspectives including the comparison of orthographies in different vernacular languages and in Latin will likely push forward the understanding of what impact the cultural context had on premodern writing. Future studies in this field will certainly profit from further development of digital historical corpora and digital methods for language analysis.

13.1 Introduction

Any piece of writing that seeks to deal with issues related to spelling standardization in a serious way cannot do so without first providing an overview of the meanings behind the term standardization. This chapter aims to do just that – to provide an overview of previous thoughts and frameworks for understanding ‘standardization’, both from a theoretical point of view, and with specific reference to recent preliminary, large-scale work in English spelling. Standardization is normally described as a process that involves several stages and follows some objectively identifiable steps, namely the selection of a norm, characterized by the minimization of variation in form, the extension of the standard to a wide range of functions, also known as elaboration of function, the codification of the forms which emerge as minimally variant, the maintenance of the standard, and the prescription of the standard to the community (Reference HaugenHaugen 1966a: 933, Reference Haugen, Ammon, Dittmar, Mattheier and Trudgill1987: 59, Reference AghaAgha 2003: 231, Reference Agha2006: 190, Reference Nevalainen, Tieken-Boon van Ostade, Hogg and DenisonNevalainen and Tieken-Boon van Ostade 2006: 274–86, Reference Milroy and MilroyMilroy and Milroy 2012 [1985a]: 23). The foundational model of description above has been modified a few times, so that the stages of standard development most commonly identified are not necessarily understood as successive, but rather they may overlap with each other or can even be cyclical (Reference LeithLeith 1983: 32, Reference Deumert, Vandenbussche, Deumert and VandenbusscheDeumert and Vandenbussche 2003: 4–7). Most recently, the understanding of standardization outlined above was updated by Agha, who sees the set of linguistic norms that come about from the process of standardization as a linguistic repertoire, which can be differentiated within the language as a socially recognized norm and linked with a framework of cultural and societal values (Reference AghaAgha 2003: 231, Reference Agha2006: 190).

All of the versions of the foundational model above have a shared, relatively straightforward and fundamental understanding of standardization. For Haugen, a standard combines the two requirements of “minimal variation in form” and “maximal variation in function” (Reference HaugenHaugen 1966a: 931). According to Reference 729MilroyMilroy (1992a: 129, Reference Milroy2001: 531), “the process of standardization works by promoting invariance or uniformity in language structure [and] consists of the imposition of uniformity upon a class of objects.” For Nevalainen and Tieken-Boon van Ostade, standardization follows two fundamental processes, maximum application, that is generality, and minimum variation (i.e. focusing ; Reference Nevalainen, Tieken-Boon van Ostade, Hogg and DenisonNevalainen and Tieken-Boon van Ostade 2006: 310). These two conditions are best fulfilled in the spelling system, but they are never likely to be met in full in any given standard or norm. Having mentioned the term norm in association with issues of standardization, it seems appropriate to also introduce a definition of the word, not least because some recent work in English spelling (Reference Archer, Kytö, Baron and RaysonArcher et al. 2015) has made explicit reference to terms like normalization. As a model or pattern, a norm is more or less codified, more or less prestigious, and it is an abstraction that emerges in a community, one which may have been formed by an authority for special purposes (Reference Locher, Strässler, Locher and SträsslerLocher and Strässler 2008: 2). According to the Oxford English Dictionary Online (Reference Simpson and ProffittSimpson and Proffitt 2000–), a norm is “a value used as a reference standard for purposes of comparison,” and, one may infer, normalization refers to the imposition of a norm as a reference standard. If the concepts of standard and norm are so closely associated, it is because a norm forms a natural basis for a standard: a linguistic norm can be selected and held up as a reference standard. However, when a linguistic norm is selected as a standard and used as a yardstick, it becomes a prestige norm and is associated with values of correctness, appropriateness and social status (see Reference BartschBartsch 1985).

The definitions above fit naturally with the understanding of standardization as a set of norms which exist in incipient stages of development and then redevelop to become less likely to vary and easier to predict. In this chapter, the said understanding of standardization represents an inevitable assumption when talking about spelling development, but is not an inescapable requirement. In principle, being in possession of a standard written system with a clear set of rules to be followed is neither a necessary nor a sufficient condition for a language to exist, if we look upon language simply as the property of a speech community with shared norms (Reference LabovLabov 1972a: 27). However, if orthography is a core element in the ideological and symbolic production of societies and languages, there is one fundamental concept that runs through the history of English, which in turn makes this language particularly suited for discussing issues about spelling standardization. Standardization is inseparably bound up with the written language, as it is in this channel that uniformity of structure is most functional and obvious (Reference HaugenHaugen 1966a: 929, Reference SamuelsSamuels 1972: 109, Reference PoussaPoussa 1982: 81, Reference Milroy and WrightMilroy 2000: 14, Reference Schaefer and SchaeferSchaefer 2006: 4, Reference Lange, Bergs and BrintonLange 2012: 995). The process of standardization, therefore, revolves especially around the act of improving on existing means of written expression, making them more uniform and complete, as well as more logically consistent and predictable (Reference Günther and StetterGörlach 1990 [1988]: 14, Reference Cheshire, Milroy, Milroy and MilroyCheshire and Milroy 1993: 3). Traditionally, the process of standardization in English spelling has been described as a unified, unidirectional narrative of linguistic development. The unidirectional perspective sees the emergence and dominance of a growing, centralized standard of written English toward which all individual spellings, and the spelling system of English as a whole, gravitate (see Reference FisherFisher 1996, especially pp. 36–83). The development of a growing standard spelling system in English was generally described as spreading irregularly over the course of two centuries (see, e.g., Reference ScraggScragg 1974: 67, Reference Salmon and LassSalmon 1999: 32, Reference Nevalainen and Raumolin-BrunbergNevalainen and Raumolin-Brunberg 2005: 42, Reference Nevalainen, Baddeley and VoesteNevalainen 2012a: 151), and it is for this reason that this chapter focuses mainly on Early Modern English. General evolutionary principles (see, e.g., Reference 701Hope and WrightHope 2000) also suggest that a complex trait like spelling, in the English language, may have evolved diachronically from simpler precursors.

The evolutionary linearity is imposed on history by a backward projection of present-day standard spelling on Early Modern English spelling, and it can be seen as a legitimate attempt to create a history for the standard language – to imagine a past for it and to determine a canon, in which normative forms are argued for and unorthodox forms rejected (Reference Milroy, Stein and van OstadeMilroy 1994: 22–23). In a way, historicizing the language becomes a necessary condition for the concept of a standard to coexist with that of a norm, intended as an elevated set of rules to be followed. The most inevitable consequence of concentrating on a unique history of standardization for Early Modern English spelling is that a history of developments in individual graphemes or groups of graphemes, some of which may perhaps have been relatively autonomous from each other and apparently counterintuitive, is made to appear as unidimensional. Such a view of the history of standardization becomes, to use Roger Lass’s words (1976: xi), “a single-minded march” toward regularity. In this sense, therefore, the concept of standardization in Early Modern English spelling is dependent on chronological dimensionality and rests on the point of view of the present-day spectator. It is especially within the remits of this point of view that one can talk about spelling standardization as a process of ‘evolution’ or ‘forward-facing development’.

13.2 Synthesis of Competing Tendencies

In contrast to the linear theoretical model of standardization expressed above, a number of scholarly voices across generations have proposed a more complicated scenario, where standardization is the result of a synthesis of competing forms and tendencies. According to this view, standardization is seen as a nodal process which sees stages of development and redevelopment through adaptation and under the pressure of sociohistorical factors (Reference SamuelsSamuels 1972: 165–70, Reference Günther and StetterGörlach 1990 [1988]: 18–24, Reference Wright, Moreno Fernández, Fuster and CalvoWright 1994: 110–13). The advent of historical sociolinguistics has reinforced the idea above, not only uncovering and describing the existence of multiple competing sociolinguistic standards, but also developing a theoretical understanding of linguistic evolution as nonlinear and nonprescriptive (see, e.g., Reference BаkhtinBakhtin 1981, Reference Bell, Llamas, Mullany and StockwellBell 2007, Reference Nevalainen, Raumolin-Brunberg, Hernández-Campoy and Conde-SilvestreNevalainen and Raumolin-Brunberg 2012, Reference Hernández-CampoyHernández-Campoy 2015, Reference NevalainenNevalainen 2015). For historical sociolinguists, the standardization of English spelling would be the result of a convergence of multiple axes of linguistic change, each initially relatively autonomous from the other. The accumulation of multiple competing iterative processes would be in itself the essence of each axis in the process of historical development toward regularity and predictability. According to this framework, the standardization of English spelling would therefore be best described as a multiple, rather than a unitary process, one that more faithfully describes “the hybrid linguistic nature of Standard English” (Reference 701Hope and WrightHope 2000: 49). The understanding of standardization as a convergence of multiple axes of linguistic change does not need to reject all internal irregularities as erroneous, but can accommodate them as “structured variation,” markers of the process of linguistic selection between competing standards (Reference Milroy and WrightMilroy 2000: 20–22). In other words, one can imagine the standardization of English spelling as the result of converging and diverging trajectories, a synergy of individual voices composing a polyphonic harmony in a chorus.

The development of structure in spelling, therefore, stems from the development of each individual spelling feature, which in turn comes out of a complex, dynamic and often unpredictable scenario. Among the linguistic elements which participated in defining English as we know it today, the development of spelling in particular is perhaps one of the most convoluted and unpredictable in the history of standardization. If any sensible generalization can be made at this point about spelling developments in English, it would be that spelling standardization as a whole is not something that we can usefully chart in a linear temporal continuum, to avoid oversimplifications and distortions of language history; instead, we can only represent aspects of spelling standardization, which can help us to glean insights into its continuous development. A more sophisticated way to understand the overall patterns of development in English would be to see orthographies within sections of the history of English as Gleichzeitigkeit der Ungleichzeitigen (see Reference BurkhardBurkhard 2002, Reference SchlöglSchlögl 2013) . This term translates to the ‘simultaneity of the nonsimultaneous’, and refers to the idea that any historical moment represents a section through the historical continuum, which reveals different time layers featuring phenomena of very different age and duration in continuous (re)development (Reference KoselleckKoselleck 1979: 132). It is perhaps due to the dynamic nature of the setting in which English develops that the spellings in the sixteenth and the seventeenth centuries in particular are in permanent development. In other words, the two requirements set for a standard language at the beginning of this chapter, maximum application and minimum variation (see Reference Nevalainen, Tieken-Boon van Ostade, Hogg and DenisonNevalainen and Tieken-Boon van Ostade 2006: 310), are only true for Early Modern English spelling if they are conceived as processes in the making.

The view of spelling standardization as a process that is never fully complete entails the idea that maximum application and minimum variation never reach a complete status, but increase in the degree of focusing, a term used by sociolinguists to refer to a high level of agreement in a language community as to what does and what does not constitute ‘the language’ at a given time (Reference TrudgillTrudgill 1986: 86). When interpreting patterns of convergence as indicators of standardization in spelling, therefore, standardization itself would be best intended as a matrix of focus rather than of fixity. In other words, individual historical spellers tended to a greater or lesser extent to conform to an idea of standard, but none of the individual texts surviving, or even a specific range of years within the sixteenth and the seventeenth centuries, can ever be said to demonstrate every characteristic of a standard. Thus, the developments toward new graphemic arrangements do not contribute to making a clear-cut set of fixed shibboleths. Instead, they give form to what the nineteenth-century scholar Alexander J. Ellis, who first described it, called a “sort of mean,” a kind of fixed magnet of spelling forms toward which printed English tended (Reference 762SmithSmith 1996: 50). The sort of mean continues to be regionally and chronologically interfered with by the “pluricentric” nature of the English language, “one whose norms are focused in different local centres, capitals, centres of economy, publishing, education and political power” (Reference Romaine and RomaineRomaine 1998: 27). If we want to look upon spelling standardization in English from a large-scale point of view, the term supralocalization may come in particularly handy for addressing the conceptual difficulties that stem from a pluricentric language (see also Chapter 21, Chapter 25 and Chapter 30, this volume).

Supralocalization refers to the geographical diffusion of linguistic features beyond their region of origin, and, in this respect, it achieves the chief goal of standardization, to reduce the amount of permissible variation while increasing in focusing (Reference Nevalainen, Tieken-Boon van Ostade, Hogg and DenisonNevalainen and Tieken-Boon van Ostade 2006: 288). If supralocalization works well as a defining term to describe the spectrum of spelling developments from a large-scale perspective, however, it alone cannot capture the full essence of spelling standardization, which stems from all of the complex range of definitions and nuances expressed in the chapter so far as a whole. Spelling developments in Early Modern English in particular developed in a way that blurs many of the distinctions between focusing, supralocalization and standardization and were never a fixed point of reference. For this reason, it would be safer, I suggest, to describe the spelling system of the Early Modern English era (i.e. the period between the sixteenth and the seventeenth century) as a system in the process of standardizing. The theoretical challenges existing to date certainly reflect the complexity of the relation between standardization and spelling, but it is not just a theoretical problem. The challenges are also caused by our little knowledge about the complex dynamics behind spelling standardization in English. It is also for this reason that, I believe, more work on large-scale developments in English spelling is needed, especially in and around the Early Modern era. The following sections provide an overview of what we know to date about the dynamics in English spelling immediately prior to and during that era, drawing on some of the most relevant recent work in the field.

13.3 A Standardizing Variety of English

The process of standardization of English was a long and complex one, and saw some fundamental developments before the introduction of printing in England in 1476. In general, the origins of a standardized form of written English have been associated with four varieties of English, identified as Types I–IV, with characteristically focused spelling conventions (see Reference Samuels and LassSamuels 1969 [1963], Reference 762SmithSmith 1996: 70; see also Chapter 21, this volume), three of which were connected with the London area (Reference Samuels and LassSamuels 1969 [1963]: 404–18, Reference Nevalainen, Baddeley and VoesteNevalainen 2012a: 133–34). Type I was the language of the Wycliffite manuscripts, some features of which appeared in vernacular medical writings from the Central Midlands in the fifteenth century (see Reference Taavitsainen and WrightTaavitsainen 2000, 2004b). Type II was the language of the documents from the Greater London area produced before 1370, and is “Essex in basis with accretions from Norfolk and Suffolk” (Reference Samuels, Benskin and SamuelsSamuels 1981: 49). Type III was the language of literary manuscripts, for example the earliest manuscripts of Geoffrey Chaucer’s Canterbury Tales, Thomas Hoccleve’s manuscripts and William Langland’s Piers Plowman, as well as London guild account records. Type IV departed “from a combination of spoken London English and certain Central Midland elements” (Reference Samuels and LassSamuels 1969 [1963]: 413–14) to become the language of the Chancery, adopted for writing royal and government documents in 1430. Type IV appeared at least two generations after Type III and was a mixture of features from Type III, with elements from the Central Midlands (Reference Samuels and LassSamuels 1969 [1963]: 411–13, Reference Benskin, Kay, Hough and WotherspoonBenskin 2004: 2; compare, however, Reference Nevalainen, Baddeley and VoesteNevalainen 2012a: 137).

Numerous scholars see the origin of a standardizing form of written English in the Chancery writings stemming from Type IV (Reference Samuels and LassSamuels 1969 [1963]: 411, Reference ScraggScragg 1974: 52, Reference RichardsonRichardson 1980: 728–29, 737, 740, Reference Heikkonen, Nevalainen and Raumolin-BrunbergHeikkonen 1996: 115–16, Reference Nevalainen, Tieken-Boon van Ostade, Hogg and DenisonNevalainen and Tieken-Boon van Ostade 2006: 274–75, Reference Corrie and MugglestoneCorrie 2012: 145–46, Reference Nevalainen, Baddeley and VoesteNevalainen 2012a: 133, 136), and suggest that the ‘Chancery standard’ was derived from the variety selected by Henry V’s Signet Office, then accepted as the Chancery norm. This presumably later developed as the supraregional standard for printers and the language community at large, possibly due to the authority of the Chancery and of the institutional endorsement (see Reference FisherFisher 1977, Reference Fisher1996). Some of the conventions derived from the Chancery documents include the spellings of the words not, but, such(e), <gh> as a velar sound in light and knight, <ig> in the word reign, and the preference for <d> (rather than <t>) in the past tense and past participle forms of weak verbs (Reference 762SmithSmith 1996: 68–69, Reference Nevalainen, Tieken-Boon van Ostade, Hogg and DenisonNevalainen and Tieken-Boon van Ostade 2006: 275–76). The written code adopted by the Chancery spread far and wide not only thanks to the circulation of documents, but also due to the frequent migrations of the clerks who had served their apprenticeships at the Chancery. While the above seems to be a widely shared agreement among most scholars, there are also contrasting voices that reject the claim that a single variety may have formed the basis for a standardized written form of English (see Reference 701Hope and WrightHope 2000: 50, Reference Benskin, Kay, Hough and WotherspoonBenskin 2004: 1–40). Instead, they suggest that an inconveniently wide range of dialects contributed to the modern standard, and that the ‘Chancery standard’ features a great deal more variation than is generally admitted. More recently, a volume edited by Reference WrightWright (2020b) has challenged the rather simplified development proposed by the orthodox origin explanations, and has pointed instead to a rather complex scenario where not only was English standardization geographically fragmented, but it was also shaped by nonlinear interactions between dialects, peoples and even foreign languages (e.g. Latin and French).

Regardless of the exact regional origins of the late medieval standardized varieties in England, it is quite clear that London was a linguistic pool from which a standard was emerging for the Early Modern English period (Reference RutkowskaRutkowska 2013a: 48). The English printing industry echoed the southern variety of English that had emerged by the late fifteenth century in the city, simply because it was there that the printing industry developed the most. Through wide-ranging trilingualism of the literate social ranks and extensive lexical borrowing, then, Latin and French spelling conventions also left their mark on the emerging standardized form of English that was being used in London. In systemic terms, the basic phonemic fit between English spelling and pronunciation, for example, was weakened by the adoption of new digraphs and grapheme/phoneme correspondences (Reference Nevalainen, Baddeley and VoesteNevalainen 2012a: 132). Some other conventions that were adopted already from the Middle English period (especially from the 1200s onward) include <ou> corresponding to Middle English /uː/ and replacing the previous <u> in word spellings like house, <qu> for earlier <cw> in queen, and <o> for earlier <u> in love. In the fifteenth century, <ie> was taken over from French to represent Middle English /eː/ and, especially at the beginning, it appeared most frequently in words of French origin (Reference GörlachGörlach 1991: 47). From the introduction of printing, London drew artisans and printing house staff from various parts of the country and from abroad, so that the London book business quickly became a melting pot of dialects and foreign languages (Reference 702Howard-HillHoward-Hill 2006: 19). Contact influence inevitably played a role in shaping the emerging modern standard that was echoed by the introduction of printing, and increased the overall complexity of the variety of spellings used in print.

Among the more recent commentators, Reference Salmon and LassSalmon (1999), Reference Nevalainen, Baddeley and VoesteNevalainen (2012a) and Reference RutkowskaRutkowska (2013a) showed evidence for interesting developments in the southern variety of English used in and around London. From the range of examples made available in their work, an increased level of focusing appears to have been in place already from the end of the sixteenth century, when attempts to systematize irregular features such as vowel length and, in some cases, vowel quality are apparent. Vowels could be doubled to mark length, and digraphs began to be used to indicate quality: for example, <ee> was used for /eː/, <oo> for /oː/ and <ea> for /ɛː/, as in seen, soon and sea. Word-final <e> could similarly indicate the length of the preceding vowel (made, side, tune), and consonant doubling a short vowel (hill ). During the sixteenth century, effort went into keeping homophones apart, and spelling words like all and awl and made and maid differently. A third improvement was to regularize the orthography borrowed from medieval French by altering the spellings, so as to reflect their supposed Latin roots. Respellings such as debt (< Middle English dette ; L dēbitum), doubt (< Middle English doute(n) ; L dubitāre), and victuals (< Middle English vitailes ; L victuālia), for example, would have represented preferable forms for those who knew Latin. Some of the most typical and widely known spelling changes that occurred over the seventeenth century, instead, are the distinction between vowels (˂i˃ and ˂u˃) and consonants (˂j˃ and ˂v˃), and the replacement of word-final ˂ie˃ with ˂y˃.

Most recently, a couple of studies have expanded on the overview provided so far, providing interesting insights into spelling developments in English from a quantitative, large-scale point of view. Reference Berg and AronoffBerg and Aronoff (2017), for example, conducted a diachronic investigation of spelling developments in four derivational suffixes, claiming that English spelling “gradually became more consistent over a period of several hundred years, starting before the advent of printers, orthoepists, or dictionary makers, presumably through the simple interaction of the members of the community of spellers, a sort of self-organizing social network” (Reference Berg and Aronoff2017: 37–38). While Berg and Aronoff ’s claims are interesting, there is another piece of research that has provided, in my view, more compelling details for the large-scale development of Early Modern English spelling. Reference Basu, Estill, Ullyot and JackakiBasu’s work (2016) goes beyond the specific case of suffixes, uses some more reliable corpus material, and makes less unorthodox claims about the process of standardization in English spelling.

13.4 Two Waves of Standardization

Basu’s quantitative investigation was conducted using an early version of Early English Books Online: Text Creation Partnership (EEBO-TCP 2015–ongoing), and identified two major waves of spelling standardization in the Early Modern English era, one running over most of the sixteenth century and another one occurring in the seventeenth century. The work analyzed the cumulative transition periods of individual word tokens to find out whether there was a general time pattern to linguistic change, and showed all new word forms that gained wide acceptance between the two centuries. The results indicated some radical differences between the two waves of development within the first two centuries of English printing, intervaled by periods of relative lull in the rate of orthographic change (Figure 13.1). The first wave of standardization, which occurred around the mid-to-late sixteenth century, appeared to be larger than the second wave, and was correlated with the period where the process of printing matured in England. The rate of words becoming established as standard detected for the end of the sixteenth century was found to be not only higher, that is, featuring a greater number of words, but also broader than the seventeenth-century wave. This pattern was interpreted to indicate that the words which shifted in spelling before and around the mid-sixteenth century were not aberrations, but, according to Basu, part of a broad, expanding set of innovations. The second wave, which occurred around the time of the Civil War, on the other hand, was seen as fitting a traditional expectation of spelling standardization in English, because it appeared as well defined and easily identifiable. The relative broadness of the two frequency waves of standardization also showed that a degree of variation remained over the course of the Early Modern era. The norms that codified as standard resulted from supralocal developments which converged to become standard usages, and formed well-defined waves of development which tended toward standardization as a fluctuating development.

Figure 13.1 Percentage of orthographic forms per year coming into widespread use

(from Reference Basu, Estill, Ullyot and JackakiBasu 2016: 194)

A corresponding barplot, drawn as part of the exploratory work in the same study (Figure 13.2), visualized the number of words undergoing a downward transition in a given year – that is to say, it plotted the rates at which orthographic forms fell out of usage. A comparison of the two plots revealed interesting patterns of development, and showed that the rate of decline for these forms lagged slightly behind the rate of acceptance for new forms. Older spellings continued to occur in relatively low numbers, even after the new forms established themselves. The spike visible for the mid-seventeenth century, instead, revealed the presence of a process where a relatively large number of remaining variant forms were finally made redundant. This outcome was, Basu argued, the end result of the growing demands for efficiency in the process of printing made by the exploding volume of printed material during the Civil War. Even more so than the seventeenth century, the sixteenth-century wave was interpreted by Basu as a very significant phase of standardization in the middle of the Early Modern era, one that has been almost completely overlooked by previous generations of scholars. The most significant observation about the first wave of standardization is the fact that its onset predates most of the prescriptive comments from early modern grammarians on the standardization of spelling. What is more, the years between the 1570s and the 1620s, during which most theoreticians published their ideas on spelling in Early Modern English, turned out to be much less unstable in terms of orthography than previously assumed. Unlike those of Reference Berg and AronoffBerg and Aronoff (2017), Basu’s findings appear to be more in line with previous statements made on Early Modern English spelling as a whole. For Reference 702Howard-HillHoward-Hill (2006: 16), printing was the cradle of change and development toward standardization. Printers manipulated spelling, and were driven by typographic pressures to serve their own business, responding to the changes which affected the printing industry between the sixteenth and the seventeenth centuries. According to Smith, the introduction of printing in particular was a key “external event” that accelerated the process of standardization and encouraged a debate around linguistic reform (Reference Smith and MugglestoneSmith 2012: 174). For Reference RutkowskaRutkowska (2013a: 62, 167, 252), combined influences from two main sources, theoreticians on the one hand and printers on the other hand, were probably responsible for shaping Early Modern English spelling in a modern direction and for influencing its speed of development. According to Reference Tyrkkö, Kopaczyk and JuckerTyrkkö (2013: 151), it has been “commonplace” to attribute the gradual standardization of English spelling in the Early Modern English period to the impact of printing technology.

Figure 13.2 Percentage of orthographic forms per year undergoing a rapid decline in usage

(from Reference Basu, Estill, Ullyot and JackakiBasu 2016: 194)

To return to the model of standardization introduced at the incipit of this chapter, the two waves of development uncovered by Basu alert us to some important qualities of Early Modern English spelling and its process of standardization. First of all, the sixteenth and the seventeenth centuries were already beyond the selection stage: by the end of the Early Modern era, a standardizing form of English had been largely selected and accepted (perhaps by those who used the ‘Chancery standard’), but codification, unlike for other European languages (for French, see, e.g., Reference Ayres-Bennett, Parry, Davis and TempleAyres-Bennett 1994: 53), was still in its infancy. The two centuries over which the Early Modern English era stretches, therefore, can be seen as a continuation of the time of elaboration (see Reference KlossKloss 1967, Reference Kloss1978, Reference FishmanFishman 2008) which had begun in the Middle English period, and had departed from the London ‘standard’ (Reference Lange, Bergs and BrintonLange 2012: 1001). While it is generally agreed that elaboration is primarily a lexical process (Reference Milroy and MilroyMilroy and Milroy 2012 [1985a]: 31), it also affected orthography, as spelling was developing to become more predictable and more functional. Despite their tendency to achieve predictability and functionality, however, variant forms are expected to continue to persist as idiosyncrasies. When taken together, both of the individual waves of spelling developments describe a process of standardization in continuous tension, regulated by forces of supralocalization. The process of standardization in Early Modern English, therefore, certainly does not fix all irregularities, but instead allows for some variation in the spelling of English even when one particular orthographic form rises to become, in practical terms, a ‘standard’.

A second interesting finding that stems from the results summarized from Basu relates to the overall conceptual understanding of standardization in Early Modern English spelling. The process of orthographic change was phased and discrete, rather than uniform and gradual, and proceeded not from chaos to order, but from coherence to coherence, through reiterative and individual changes. Moreover, standardization was an ongoing, slow development, which never fully completed by the end of the seventeenth century. For this reason, as mentioned earlier on, diachronic spelling developments in Early Modern English, which tend to become a more focused set of norms, should be seen as elements of an incipient and ongoing process of standardization, rather than complete, stand-alone elements of a standard. In light of the quantitative overview given in recent research, a distinction between regularization, seen as a univocal and unidirectional development toward standardization, and standardization, seen more as an umbrella term for encompassing multiple stages of regularization, should be retained also in future large-scale work on diachronic spelling development. Without a doubt, the findings made available by Basu from recent, preliminary work on large-scale developments in Early Modern English spelling represent a useful and inspirational point of departure for future work in the field. At the same time, however, much as for Berg and Aronoff’s study, there are elements from Basu’s work that need to be taken with caution.

In particular, some statements more than others, not least the bold claims about the wholesale influence of the printing press on large-scale spelling developments, need to be corroborated with more substantial linguistic evidence than was made available in Basu’s chapter-long work. Previous statements in Reference Basu, Estill, Ullyot and JackakiBasu’s study (2016: 183–84), for example, suggest that graphemes form the basic units for the process of standardization in Early Modern English, which means that a more extensive quantitative focus on graphemes is a desideratum in order to glean an understanding of the dynamics involved in the standardization of English spelling. Moreover, the influence of theoreticians, schoolmasters, authors and readers, as well as printers, on the development of English spelling remains, to date, a widely debated, unsettled topic (see Reference ScraggScragg 1974: 52–87, Reference Salmon and LassSalmon 1999: 32, Reference Nevalainen, Tieken-Boon van Ostade, Hogg and DenisonNevalainen and Tieken-Boon van Ostade 2006: 290, Reference Percy, Bergs and BrintonPercy 2012: 1008). For Reference Rutkowska and RussiRutkowska (2016: 187; see also Chapter 22, this volume), “the question of the extent to which theoreticians, on the one hand, and printers, on the other, have impacted on the regularization of Early Modern English spelling is likely to remain a chicken-and-egg problem for quite a long time.” In my opinion, this is because systematic, large-scale studies are necessary to give more meaningful answers to big questions like the one posed above. Recent work that seeks to fill this gap in an informed way (see, e.g., Reference Condorelli and CondorelliCondorelli 2020b, Reference Condorelli2021a, Reference Condorelli2021b, Reference Condorelli2022b) needs a compendium of knowledge about the elements mentioned above. To this end, the following section describes the role and relevance of theoreticians, schoolmasters, authors and readers in the Early Modern English era. Another chapter in the present Handbook (see Chapter 23, this volume) gives more insights into printing and printers in England.

13.5 A Multiparty Affair

Prescriptivism in the English language is a practice almost as old as the Early Modern English era itself. The first comments concerning the need for a spelling reform in English appeared around the middle of the sixteenth century, as a response to John Cheke’s and Thomas Smith’s interest in the pronunciation of ancient Greek (Reference ScraggScragg 1974: 59). While Cheke did not explicitly formulate any reform proposals, his translation of fragments of the Gospel of Saint Matthew was an implicit statement in this sense (Reference DobsonDobson 1957, I: 43, Reference Salmon and LassSalmon 1999: 20). The most remarkable examples of spellings used by Cheke include the doubling of vocalic spelling and the use of word-final <e> to indicate vowel length (both as separate features and combined). The occasional doubling of consonant graphemes to indicate shortness of the preceding vowel also constituted a relatively regular feature in Cheke’s translation (e.g. Godd, see Reference DobsonDobson 1957, I: 44). One of the first explicit improvements proposed for English spelling, instead, was occasioned by Thomas Smith, in his Latin treatise published in 1568. Smith’s thoughts were echoed in English by John Hart, who was among the most innovative and productive advocates of a new spelling system based on pronunciation. John Hart maintained that it would be “much more easie and readie […] for the writer and printer” (1955 [1569], fol. 4) to spell English if his suggestions were followed. Hart pointed out that “we should not neede to vse aboue the two thirds or three quarters at most, of the letters which we are nowe constreyned to vse, and to saue the one third, or at least the one quarter, of the paper, ynke, and time which we now spend superfluously in writing and printing” (fol. 5).

Among those that followed Hart, Reference 745PuttenhamPuttenham (1968 [1589]: 120–21) actively promulgated rules and patterns for spelling the English language and warned his audience against adopting the speech and spelling habits of lower-class, uneducated people and countrymen (Reference Salmon and LassSalmon 1999: 17). Even more so than Puttenham, Bullokar was bitterly critical of contemporary usage among the linguistic philosophers. For Bullokar, spelling conventions had become unpredictable because of the absence of rigorous norms which could maintain correct spelling forms. “For lack of true orthography,” he said, “our writing in Inglish hath altered in euery age, yea since printing began” (Reference BullokarBullokar 1968 [1580b]: 2). Unlike his predecessors and some of his contemporaries, William Bullokar heralded a line of language theoreticians who were active between the mid-sixteenth century and the early seventeenth century and who were opposed to a radical change of the English spelling system. Bullokar’s approach reflected the growing need for control over orthography to allow for intercommunication, in a way that would not override the limitations of technology, while also avoiding complete subversion of the inherited English spelling system.

The complexity of the inconsistencies and variations in the English spelling system encouraged even further attempts at spelling reform, which were fueled by a strong interest in the Classical languages, and appreciation of their structure, literature, as well as fixed orthography (Reference GörlachGörlach 1991: 50). Admittedly, even the earliest linguistic theoreticians had acknowledged the power and tenacity of custom as relevant for influencing English spelling. Hart had described custom as “any peoples maner of doings,” and had affirmed that he was writing for the “good perswasion of a common commoditie” to those not “obstinate in their custome” (Reference HartHart 1569: 3–4). It was Richard Mulcaster, however, who was most receptive to the power and importance of custom, with all its imperfections, as a basis for shaping English spelling: “The vse & custom of our cuntrie, hath alredie chosen a kinde of penning, wherein she hath set down hir relligion, hir lawes, hir priuat and publik dealings” (Reference MulcasterMulcaster 1582: N1 v). Mulcaster tended to juxtapose “custom” and “reason” to argue that any prescriptive or reasoned philosophical argument on spelling had to take prevalent scribal and printing practices into account, although he did emphasize writing as the true measure of “custom” compared to the often error-strewn process of printing (see also Chapter 5, this volume, for more discussion on the differences in views between Hart and Mulcaster).

As well as a language theorist, Mulcaster was an experienced schoolmaster, with an interest in designing, codifying and promulgating a system of spelling rules for the use of teachers and, perhaps indirectly, even pupils (Reference ScraggScragg 1974: 62). As evidenced with Mulcaster’s example, the relationship between theoreticians and schoolmasters at the end of the sixteenth century was probably a close one. From Mulcaster’s time onward, spelling books, grammars and dictionaries became increasingly common in the Early Modern English world, to such an extent that, according to some scholars, their prescriptivist attitude also affected the work of printers. According to Reference BrengelmanBrengelman (1980: 333–34), schoolmasters sought and followed recommendations from early theoreticians, and most of the spelling principles that we use nowadays were in fact the result of a collaboration between the two. A similar idea resonates in Reference Salmon and LassSalmon (1999: 32, 34), who maintains that schoolbooks, and especially Mulcaster’s Elementarie, most likely affected printers, presumably through schooling. Examples of late sixteenth-century and early seventeenth-century schoolbooks which may also have been influential include those by Reference ClementClement (1587), Reference CooteCoote (1596), Reference HumeHume (1617), Reference EvansEvans (1621), Reference 660BrooksbankBrooksbank (1651) and Reference HodgesHodges (1653). For Reference RutkowskaRutkowska (2013a: 164), instead, lexicographers’ practice may have gradually had a role in crystallizing some spelling principles. Key titles for dictionaries and word lists include the work written by Huloet as early as 1552, later republished in a substantially expanded edition by Reference Huloet and HigginsHuloet and Higgins (1572). The earliest English dictionary was written by Reference CawdreyCawdrey (1604), a title which was likely later taken as a reference point by Reference ButlerButler (1633), Reference HodgesHodges (1644, 1649, 1653) and Reference WhartonWharton (1654).

If prescriptivism was a growing force, the figure of the author occupied a seemingly restricted, but nonetheless still important place in the thriving culture of the Early Modern English era. During the sixteenth and the seventeenth centuries, the author gradually joined the printer as a player in the story of the printing world, and was perceived as an important component within a broader system that created and distributed writing (Reference Wall and KinneyWall 2000: 69). The emerging figure of the professional author in Early Modern English is inevitably bound not only with that of the printer, but also that of the reader. Even before the introduction of printing in England in 1476, writers imagined readers as possessing the authority to change the text, turn a page or move away from a text, and often established a discourse that emphasized these and other modes of readers’ participation. Along with recognizing readers’ capabilities, authors also acknowledged both the potential and the threat offered by this participation to support or undermine authors’ own power. In other words, the attention paid to writerly authority is incomplete without attending to its complement, readerly authority. One cannot understand Early Modern English authors without also taking into account Early Modern English readers, their relations to each other and the meaningful roles played by each party – both through the ways in which authors anticipated readers’ participation, and as readers effected it (Reference BlattBlatt 2018: 2).

In practical terms, readers presented the potential for productive partnership with authors through their participatory engagement with texts, even as they concurrently threatened to disrupt the printing schedule. Accordingly, authors in early modern England focused on how readers could help or hinder through participatory reading practices, in order to maintain and develop their own status and engage readers in their projects (Reference BlattBlatt 2018: 9–10). Authors were deeply interested in shaping the practice of readers’ corrections, seeking to guide how and what readers responded to before readers ever set pen to parchment. In a way, authors explored what readers might become to them by anticipating readers’ enthusiasm for textual participation, and viewed readers as possessing a growing power to contribute in sophisticated ways to the literary language (Reference BlattBlatt 2018: 13, 19). Readers could in turn support or disrupt authors’ plans for their texts and, by extension, also those of printers. Additionally, readers could read in ways that supported beneficial interpretations of a text, or not read in a productive way at all, limiting the success of a book. Thus, readers were, in their actions and in their potential, figures whose status provoked the ongoing interest and concern of authors and printers. Both readers and authors in the Early Modern English era, then, were in turn subjected to what Reference Milroy and MilroyMilroy and Milroy (2012 [1985a]: vii) call “the complaint tradition,” a culture of tension toward the maintenance of a standardized written language. The surge of prescriptive influences between the sixteenth and the seventeenth centuries was only a prelude to the massive legislation and prescription which occurred during the second half of the eighteenth century (Reference Stein, Stein and Tieken‒Boon van OstadeStein 1994: 5). What Reference BlankBlank (1996: 9) terms “prescriptivism” in the early modern era was “diagnostic in its methods and its aims” and hence, in terms of the framework provided by Reference HaugenHaugen (1966a), it can be located before the codification phase. Complaints about English that emerged from the very onset of printing in England often focused on the inadequacy of the vernacular, and did not establish a model to follow. Nevertheless, they still put pressures and expectations on individual authors and readers respectively (see Reference DobsonDobson 1955: 27, Reference Görlach and LassGörlach 1999: 475, Reference BrackmannBrackmann 2012: 56–57, Reference Beal, Kytö and PahtaBeal 2016: 309), a matter which takes us back, in a circular way, to the influence of theoreticians and schoolmasters.

Clearly, the context where Early Modern English spelling standardized was one of synergy and contrast between “competing magnets of prestige” (Reference 762SmithSmith 1996: 65). On the one hand, these parties most likely worked in conjunction with natural, language-internal processes of competition and self-organization (see, e.g., Reference Kroch, Beals, Denton, Knippen, Melnar, Suzuki and ZeinfeldKroch 1994, Reference EhalaEhala 1996, Reference Rocha, Gertrudis, Salthe and DelposRocha 1998, Reference De Boer, Tallerman and Gibsonde Boer 2011), and operated in synchrony toward shaping the hybrid features of a more focused variety of English (Reference 701Hope and WrightHope 2000: 52). On the other hand, if taken individually, each of the external parties identified above acted as impulses or as limiting factors toward regularization and standardization. Each of the parties may in principle have hindered, weakened, slowed down or, on the contrary, reinforced, accelerated or even helped standardization, for reasons existing within the nature of orthography itself. Orthography is a practice which is bound up with other practices to do with literacy, which are themselves embedded in the social, economic, cultural and technological layers of the society where the spelling has grown. For Sebba, “any explanatory account of orthography-as-practice must be sociocultural in nature” (Reference SebbaSebba 2007: 14) and “the practices involving literacy in which a community engages are inevitably related to the type of orthography which will emerge as one of the technologies underpinning those practices” (Reference SebbaSebba 2007: 23–24). These statements somewhat echo those made by scholars like Nevalainen, who has claimed that the standardization of English spelling was in fact “a multi-party affair” among language theoreticians, schoolmasters, authors, readers and printers (Reference Nevalainen, Baddeley and VoesteNevalainen 2012a: 156).

13.6 Conclusion

This chapter has introduced readers to the concept of spelling standardization, offering an overview about the ways in which spelling standardization occurred, the agents behind the modern-like developments in historical spelling, and the chronology of the process of development in historical English. The chapter has departed from the idea that historical spelling represents one of the most complex facets of linguistic standardization, and one where disagreements exist about its overall process of development. The contribution has moved on to discuss the idea that standardization in English spelling was, for some scholars, an intralinguistic, spontaneous process of self-organization, and for others a multiparty affair that involved authors, readers, the printing press and linguistic commentators of the time. It is hoped that the present chapter will provide future scholars with useful information for a more informed approach to exploring issues and questions that remain unresolved in English spelling, following, for example, the quantitative steps laid out in Reference CondorelliCondorelli (2022b). With the experimentation of new tools to approach and understand the study of orthography, some of the existing questions can now be explored in a much more systematic way than before, giving researchers confidence to provide more meaningful answers. While implementing and testing quantitative methods geared specifically for studying something as complex as historical spelling are anything but easy tasks, they undoubtedly offer intriguing and not infrequently surprising insights into patterns that we would not be able to see with the naked eye.