As early as the eighteenth century, grammars of the English language included judgments on the grammatical status of historical developments in the syntax of the inflected forms of the interrogative and personal pronouns. At present, the use of who for whom as the so-called object form of the relative and interrogative pronoun is generally accepted, at least in the United States, as standard colloquial speech. Similarly, the defense of It is me against It is I has a lengthy history. Finally, constructions like Him and me are staying, while not accepted as standard English, have been acknowledged as forming as much a part of a natural and consistent linguistic system (so-called vulgar English) as their socially more acceptable kin. The changing syntax of the pronouns is, of course, basically a historical phenomenon. Because of the normative approach and conservatism of the schools, as well as to the strong formative influence of literary tradition, many educated speakers of American English have, as more or less distinct styles of speech, the different syntactic subsystems characterizing these usages.

1. Introduction. In the general theory of linguistics, dependency theory competes for the same role as immediate-constituent (IC) theory: each formalizes a view of tactic relations among elementary units in utterances. In each theory, a finite apparatus is taken to characterize (or generate) a set of utterances that may be infinite. In addition, each theory specifies a structure, or perhaps alternative structures, for each utterance it generates.

1. In 1956, Greenberg proposed eight measures of linguistic diversity designed to determine the possibilities of communication among the population of some delimited area. Taking note of the considerable variance between areas of the world in their linguistic diversity, Greenberg's measures allow us to quantify this diversity on a continuum ranging from complete diversity, such that no two people speak a mutually intelligible language or dialect, to the other extreme, where all inhabitants share a common tongue. By going beyond mere verbal descriptions of an area's linguistic diversity, these quantitative measures enable us to correlate language usage with social, economic, political, and geographic factors.

It is well known that in a number of European languages the perfect form of the verb employs an auxiliary identical with the verb used for predicating possession—e.g. English I have loved, French j'ai aimé, Portuguese eu tenho amado, Modern Greek ėkhō . From discussions by Vendryes and by Benveniste it has also become clear that in transitive sentences the similarity between perfect and ‘possessive’ expressions is not confined to cases involving an auxiliary ‘to have’.

Linguists who conceive of their science as a discipline which collects utterances and classifies their parts often pride themselves on their freedom from mentalism. But freedom from mentalism is an inherent feature of the taxonomic conception of linguistics, for, according to this conception, a linguist starts his investigation with observable physical events and at no stage imports anything else.

On the basis of Varro's etymology ‘uncia ab uno’ (LL 5.171), it has been said that Lat. uncia ‘ of anything, esp. as a weight ( of a pound) and monetary unit ( of an as)‘ represents an earlier *oiniciā. There are two possibilities of development: *oiniciā > *ōnciā > *ŏnciā > uncia, or *oiniciā > *ūnciā > uncia. The first requires a change known only in nōn < noenum (provided this etymology of nōn is accepted) and three regular changes—syncope, shortening of a long vowel before nasal plus consonant, and ŏ > ŭ before /η/. The second is not debatable, for oi > ū and shortening of the long vowel are both regular; given the etymon, therefore, it is the preferable development.

[The graphs 〈etuvanoinu〉 and 〈etokoinu〉 on the bronze tablet of Idalium do not contain optatives duwánoi and , but aorist indicatives éduwan and édōk'. Éduwan comes from *édoan for édon in the same way that éthijan comes from *éthean for éthen. The infinitive dowénai is scarcely dow-énai: evidence outside Italic for a root form dow- is not convincing. Either it is do-wénai, with a suffix otherwise unknown in Greek, or a development, partly analogic, from *do-énai. 〈oinu〉 is perhaps oîn nu, containing a particle cognate with Attic oûn; PGk. *óen could account for both, since a sound law PGk. *-en > Cypr. -in can be set up. The Cypriote thematic active infinitive ending is -ēn, not -en.]

Every Serbo-Croatian verb form consists of the constituents listed in Table l. While prefixes are optional, every verb form has a stem and at least one inflectional element. The stem consists of at least a root. Every verb paradigm has a thematic vowel in some or all of its forms, though some forms of some paradigms are without this. The remaining stem constituents are optional. The constituents always occur in the order given in Table 1.

One of the factors preventing the wider acceptance of the laryngeal theory of Proto-Indo-European phonology may be that it is rarely presented in a consistent manner. Those who accept the validity of the theory generally work from it as a basis rather than towards it as the thing to be explained. Hence an opponent of the theory is apt to be presented with arguments of several different sorts. One such argument is the pragmatic one that frequently runs : The problem of (say) Brugmann's Law, long unsolved, is amenable to solution by means of a laryngeal. One may object on the one hand that every problem is amenable to ‘solution’ provided enough new entities are introduced, and on the other that the problem in question is perhaps amenable to solution in another way. The laryngealist will retort that another, seemingly quite unrelated problem, generally in another language, is also solved by applying a laryngeal solution. Since solving two problems with one theory is better than solving them with two, this argument from application frequently serves to silence, if not to convince, the opposition.

Comparison of ten Rhaeto-Romance dialects reveals that, in addition to Western Romance consonant lenition, the phonological innovations which characterize Proto-Rhaeto-Romance as against its source, Proto-Italo-Western, are the following: phonemic vowel length, posttonic syllabic consonants, a phoneme , a fronted /a/, the conditioned split of PItW , and a central rounded vowel series. From these innovations it is clear that the RR dialects are so closely related to French as to make attractive the detailed comparison of the two stocks. In this article it will be assumed that a linguistic stock x, the ancestral form of French and Rhaeto-Romance, became distinct after the beginning of the Christian era, that the six innovations arose in this stock to the exclusion of the rest of the Romance-speaking world, and that the Germanic invasions ultimately broke its unity by driving the two substocks territorially apart and out of communication with each other.

Antoine Thomas first suggested in 1896 that Fr. lingue ‘ling, Lota molva (a fish)’ is of Germanic origin. His discussion of the word, published first in Romania and repeated shortly thereafter in Thomas’ Essais de philologie française, was a sober, concise, but thorough treatment which included all pertinent evidence that has been published on the subject. Later discussions have added nothing. They have either repeated fragments of the evidence cited by Thomas, usually omitting portions of the remainder which tended to balance those they used, or have adduced material which is irrelevant. Virtually nothing will be added here but evidence which would be irrelevant were it not necessary to refute erroneous proposals made later. Thomas' cautious conclusions were hardly more specific than indicated above. He stated that the French word must have been borrowed from Germanic, but could have been taken from English, Flemish, Dutch, or one of the Scandinavian languages. This proposal is clearly preferable to the only explanation previously suggested, derivation from Lat. lingua ‘tongue’ because of the physical appearance of the fish; as far as it goes, it has apparently been accepted without question. The subsequent history of the etymology, however, has been one of confusion and inaccuracy, with the result that its present state constitutes a regression from that at the time of Thomas' article.

Traditionally, the Old Church Slavonic verb is considered to be capable of forming three types of aorist. The so-called root aorist is presumably formed by adding personal endings directly to the stem (Diels §114.1), as in the forms of the verb ‘to fall with the basic stem in pad- : 1st sg. 2d and 3d sg. pade, 1st du. padově, 2d du. padeta, 3d du. padete, 1st pl. 2d pl. padete, 3d pl. The so-called s-aorist (Diels §114.2) is presumably formed by adding the element s (which under certain conditions alternates with x) directly to the stem, as in the respective forms of ved- ‘lead’ and rek- ‘speak’: , vede, věsově, věsta, věste, , věste, , and , reče, rěxově, rěsta, rěste, , rěste, . The so-called x-aorist (Diels §114.3) is formed by adding the element ox, as in the x-aorist forms of ved- : , vede, vedoxově, vedosta, vedoste, , vedoste, and . In addition to automatic changes (such as x to š before a front vowel), it must be noted that the second and third persons singular are always formed on the root-aorist pattern; and that the third person plural ends in in the root aorist but in in the s- and x-aorists, as in the forms (root aor.), (s-aor.), and (x-aor.). The distribution of the various types of aorist among stem classes is the following: class 1.1 (cf. Diels §105–6) has all three types in stems that end in an obstruent, as in the forms and stems of class 1.1 that end in a sonorant take s-aorist only; stems of class 2.1 take either the root aorist or the s-aorist; stems of class 2.2 and 3–6 take the s-aorist; and stems of class 7 (irregular) require individual disposition.

1. Since Meyer-Lübke, it has been a commonplace of Romance phonology that Proto-Romance voiceless stop consonants have, by and large, been preserved intact in word-initial position throughout the Romance territory, whereas intervocalically they have undergone lenition, becoming voiced (and in some regions also fricative and eventually dropping). The spirants /f/ and /s/ (for which there were no voiced counterparts in Proto-Romance) have likewise become voiced intervocalically. Similarly, voiced stops are normally preserved initially, but have remained intervocalically as plosives (with the exception of /-b-/> /-v-/) only in East Romance, and have become fricatives or have dropped in the West. It is convenient to refer to these West Romance developments by the general term ‘the West Romance sound-shift’, as exemplified in Table 1 (with ERomance examples from Italian and WRomance from Provençal).

A number of linguists have attempted to establish a traditional phonemic system for Modern Faroese, but they have uniformly come to grief over the vowel segments. Since there are only a very few minor problems involving the phonetic facts, it is clear that the difficulties lie not in our knowledge of Faroese but rather in the linguists themselves and most particularly in their insistence on honoring the invariance condition. The invariance condition is the demand that every occurrence of [x1] be considered an allophone of /x/ and that [x1 x2 x3…], allophones of /x/, share at least the defining features implied in /x/. This restriction does not allow the assignment of some [x1] to /x/, some to /y/, etc., nor may [x y z …] be derived from /x/ by the change of some feature implied in /x/.