We have already mentioned in this volume the importance of the historical toponomastics approach to the historical-linguistic reconstruction of place names. Toponyms, in turn, can provide indispensable clues for the study of the diachronic developmental stages of the languages to which the analysed toponyms belong to, and of the proto-languages from which those languages derive. Indeed, through the reconstruction of place names it is possible to postulate settlement dynamics and population movements that are not recorded by historical sources because they occurred prior to the invention of writing, and therefore in prehistoric times. Toponyms can even be used as clues to decipher an ancient language, as in the case of Mycenaean Greek and Linear B, explained in Chapter 2. The focus of this chapter is to provide an outline of how toponymy can help us to uncover archaic linguistic features of proto-languages and even of postulated pre-languages, which are, by definition, unattested and reconstructed. Pre-languages are languages that were lost a long time before proto-languages began to differentiate into historical, attested languages. In the field of historical toponomastics, this is accomplished mainly through etymological reconstructions of ancient place names.

The essential question that governs all diachronic reconstructions in historical linguistics is: ‘what people were there before, and what language(s) did they speak?’. Unavoidably, this question, at the etymological level, must stop at the stage of the reconstruction of proto-forms and roots. These are the basic, most remote, and abstract components of a proto-language, and they are therefore the last possibly reconstructable lexical and morphological items.

A number of issues affect a comprehensive historical-toponomastic approach to ancient toponymy. First, how far back in time can we go with our reconstructions? At some stage, we will run out of written documentation or artefacts that may substantiate our findings. That is, once we go far enough back in time, namely in prehistoric times, we will not have any historical records to refer to, because writing was not invented yet at that developmental stage of mankind. We can then only hypothesise and not discuss our results with certainty. At the same time, the goal of our analysis is to linguistically recover the most ancient proto-forms and roots of place names. This means that, in specific cases, we need to postulate the existence of a pre-language that predates a proto-language.

These issues open a theoretical debate on the notions of ‘pre-language’ and ‘proto-language’, which will be briefly discussed below.

Ancient stems, derived from the reconstruction of place names, can be used to recover the general lexicon of a proto-language. These roots are ‘ancient’ by definition, because they represent the most remote reconstructable lexical units of a proto-language, to the point of being relatively theoretical and abstract linguistic elements, always unattested, in themselves, and reconstructed. Indeed, it is because they are basic morphological components of prehistoric toponyms that they are so important in the context of historical toponomastics. This is because they not only allow the toponymist to reconstruct the remote forms of place names, but they also enable linguists to connect them with the most ancient developmental stages of prehistoric proto-languages. As mentioned in earlier chapters, extremely ancient toponyms were named prevalently after primary goods and aspects of everyday life (e.g., water, edible crops, animals, shelters, fire, stone, dangers) by prehistoric peoples, or referred to the hydro-geo-morphology of the landscape (e.g., flowing waters, mountains, cliffs, caves, fords, swamps, glades). Place names functioned as intangible aids to the earliest settlers to establish a mental and ideal map of their territories, and to help them to orientate around the landscape, in what can be defined as ‘primordial natural navigation’.

4.1 Pre-Languages and Proto-Languages

At the level of definition, a proto-language is an unattested, remote, abstract, and hypothetical language spoken in the distant past and in the absence of writing or of any other form of historical documentation, by prehistoric individuals belonging to the genus Homo. The proto-language is reconstructed, through the comparative method, on the basis of the attested (still living or already dead) and therefore historically documented languages which derive from it. A proto-language represents a linguistic stage predating the differentiation of the languages which have been generated from it when the communities speaking that proto-language became geographically separated and their ways of speaking, through the normal process of language change (cf. Chapter 2), evolved into different languages that are historically attested. All these (new) languages belong to the same language family which, generally, assumes the name of the proto-language. An example is English, which is an Indo-European language derived from the (Proto-)Germanic branch of Indo-European. The languages in a language family share lexical items (with their related meanings) and morphological and syntactic structures. In Chapter 2, we showed how, by applying the comparative method and through a reconstruction of sound laws, it is possible to reconstruct the proto-language from which all the languages belonging to a language family derive. A classic example of this is Proto-Indo-European, the proto-language at the origins of the so-called Indo-European languages. It has been reconstructed on the basis of the comparison and analysis of the attested Indo-European languages themselves, both still living (e.g., Albanian, Lithuanian, Italian, and Modern Greek) and already dead (e.g., Ancient Greek, Latin, Gothic). Through the analysis of the Indo-European languages and the recovery of the original proto-language, moreover, we can reconstruct ‘second-generation proto-languages’, like Proto-Germanic, derived from Proto-Indo-European and at the origins of the Indo-European Germanic languages, and Proto-Celtic, derived from Proto-Indo-European and at the origins of the Indo-European Celtic languages, and so on (Figure 4.1). Moreover, among Indo-European languages, we can make a distinction between a ‘first stage’ of attested (historical) Indo-European languages, at the level of ‘seniority’, for example, Sanskrit, Ancient Greek, and Latin, and a ‘second stage’ of attested (historical) Indo-European languages, derived from some of the ‘first stage’ Indo-European languages, like, for example, Modern Greek and Romance (Neo-Latin) languages like Italian, French, Romanian, Spanish, and Portuguese.

Figure 4.1 A family tree showing some of the 144 languages within the Indo-European language family

A pre-language, conversely, is (at the theoretical level) a languageFootnote ¹ spoken in a specific territory by peoples who pre-existed the new settlers (i.e., speakers of a new language). When a pre-language and a proto-language, which are both unattested by definition, are involved in language contact, that means that the dynamics between an old population and a new one happened in very remote times. The new settlers, by occupying the same territory of the old inhabitants at a later time, replace the previous peoples and their language. They are, indeed, bearers of a new language (or family of languages) not related to and incompatible with the language(s) of the previous inhabitants. A pre-language, therefore, by definition comes before a proto-language. In the case that the speakers of the proto-language completely erase linguistically the pre-language layer and erase ethnically the speakers of the pre-language, in the absence of any written record, the pre-language becomes lost forever.

To be clearer, a proto-language can be reconstructed, while a pre-language is purely ‘theoretical’ and, unfortunately, we cannot reconstruct it. We know it existed, because we have evidence (as documented by the material culture of the people speaking it and, literally, by their bones and genes) of populations inhabiting specific areas before the arrival of the proto-language speakers. These people possibly lived in these areas for thousands of years, but their unattested and unrecorded language or languages have been irremediably lost. Using the Indo-European context as an example, we know that Europe, before the possible arrival of the Indo-Europeans (who were the linguistic bearers of Proto-Indo-European), was inhabited for tens of thousands of years by other peoples belonging to the genus Homo. These people surely were speaking a variety of (Pre-Indo-European) languages, which were all erased with the arrival of the Indo-Europeans, and which were superseded by Proto-Indo-European.

Generally, a proto-language and a pre-language, even if spoken in the same area, are not considered to be connected or related to each other. ‘Proto-Indo-European’ therefore indicates what possibly was the common prehistoric Indo-European (proto-)language before branching out to the ‘second-generation’ proto-languages derived from it and in the attested historical Indo-European languages. ‘Pre-Indo-European’, conversely, despite containing the notion of ‘Indo-European’ in its name, was probably a non-Indo-European language (or set of languages) in its intrinsic nature, and pre- in its name means exclusively that it was pre-existing and predating (Proto-)Indo-European.

In current scholarship, as mentioned earlier in this book, a proto-language usually consists of a collection of reconstructed roots and proto-forms which have been recovered by analysing cognates belonging to all the languages derived from it through the application of the comparative method (cf. Chapter 2). This method also allows us to understand which branches of a language family developed earlier or later, over time. While a proto-language can be reconstructed with the sound laws regulating its development into the ‘second-generation’ proto-languages and the attested historical languages derived from it (as seen in Chapter 2) through the comparative method, a pre-language can only be hypothesised. Indeed, the existence of very remote civilisations speaking pre-languages can only be inferred through the analysis of physical evidence from the fields of (prehistoric) archaeology and paleo-anthropology. For example, the existence of a Pre-Indo-European culture seems widely attested by the numerous prehistoric sculptures of the pregnant Venus-type found across Europe by archaeologists like Reference GimbutasGimbutas (1974; Reference Gimbutas1989; Reference Gimbutas1991; Reference Gimbutas1999). Those artefacts seem to attest to the existence of a strongly matriarchal society (or societies)Footnote ² occupying the Old Continent (i.e., Europe) way before the possible arrival of the Indo-Europeans (and the Indo-European languages). This society (or societies) would have had its own language (or languages), which was then supplanted by Proto-Indo-European and the derived Indo-European languages.

The only possible way to reconstruct plausible ‘linguistic fossils’ of pre-languages is the analysis of presumably prehistoric place names. Toponyms, indeed, could indirectly preserve some traces of the languages spoken before the arrival of speakers of a proto-language. This is because, as we have already mentioned, ancient place names tend to be stable over time. Even with possible morphological modifications, they are not (drastically) changed by new people, especially in their roots. This applies not only to Europe, but, more or less, to linguistic contexts all over the world, from prehistoric ages, and also in more recent times. This is also true in aboriginal and/or Indigenous environments, within communities inhabiting remote territories, and in isolated areas where writing is not used or has not been developed yet.

To sum up, a pre-language is generally considered unrelated to a proto-language that replaced it and was spoken before the arrival of the speakers of that proto-language. The existence of a pre-language can only be hypothesised due to the massive lack, or total absence, of documented evidence. However, toponyms may be the missing linguistic links we need to prove its existence and to ideally give back the voice to its very ancient speakers. This is only a simplified representation, but it emblematically underlines the theoretical differences between the notions of ‘pre-language’ and ‘proto-language’.

Moreover, it is possible that, because of cultural contact between the speakers of a pre-language and the speakers of a proto-language, some lexemes from the former were transferred into the linguistic system of the latter. Those lexemes, despite being adapted to the phonetic, phonological, and morphological standards of the ‘new’ proto-language, may retain some of their original features. An analysis of these possible linguistic ‘anomalies’ can generate the hypothesis according to which these ‘contact features’ are not ‘native’ to the proto-language, but derive from the pre-language that pre-existed it.

Later in this chapter, we will discuss three toponymic systems that could shed light on the relations between the notions of ‘proto-language’ and ‘pre-language’ and Toponymy. These examples will postulate the existence of roots belonging to a Pre-Indo-European stratum, which survived the possible arrival of the Indo-Europeans and which were reused and re-functionalisedFootnote ³ in the Indo-European phonetic, phonological, and morphological systems.

Before we go on, it is worth briefly outlining some theoretical positions about the notions of ‘Pre-Indo-European’ and ‘Proto-Indo-European’. A discussion on pre-languages and proto-languages, indeed, can only be conducted in the context of well-researched and documented language families and, to date, the best candidate is still the Indo-European context. Indo-European languages are very well-researched, but that does not imply that all scholars are in agreement about their origins. We will briefly introduce here, therefore, some well-known theories on who were the speakers of Proto-Indo-European and their possible relation to Pre-Indo-European civilisations. Some of these positions are dated and obsolete, or have been the object of quite heated debates. However, in order to provide our readers with a multifaceted perspective on the issue of the Indo-European origins, we feel it is worth summarising them here. We will, convergently, also focus on the theories in which the notion of ‘pre-language’ can be relevant to toponymy. The aim of this chapter is not to provide an exhaustive discussion on these theories (which is beyond the scope of this book), but to highlight the role that toponymy plays in the studies of Indo-European languages, and to show how some European place names are possible evidence of the presumable existence of Pre-Indo-European language(s) and can be gateways to a better (and unbiased) understanding of prehistoric Europe.

4.2 Giacomo Devoto’s ‘Mediterranean Theory’

In language-contact contexts, when pre-existing peoples are subjugated by other (new) ethnic and linguistic groups, or when people speaking different languages (and/or languages belonging to different language families) merge into one community, one of the results of the process is that the descendants of both groups generally end up speaking a single language. While the resulting language will most likely be a variant/development of the language spoken by the new people, it will probably also include features belonging to the old language. An Indigenous or local language that provides features (lexical, phonetic, syntactic, etc.) to the language of the people who invade and impose their language on the local population is referred to as ‘linguistic substrate’. This language leaves discernible traces in the new language spoken by all the ‘merged’ people in a specific area. In the European prehistoric context, with the possible arrival of the Indo-Europeans, the Pre-Indo-Europeans were overrun by the newcomers. The end result of this process was a replacement of the Pre-Indo-European languages by the new Proto-Indo-European language (or already differentiated (proto-)languages spoken by the Indo-Europeans). However, despite the linguistic change, some features of the previously spoken language(s) could have been adopted by, or incorporated into, the new language(s). The hypothetical existence of a Pre-Indo-European linguistic substrate seems to emerge from some very ancient European place names.

Giacomo Devoto hypothesised that the Indo-Europeans, particularly the (proto-) speakers of Ancient Greek and Latin, came from northern and/or central Europe (Reference DevotoDevoto, 1962). They then would have settled in the Mediterranean area that spans across three modern continents (Europe, northern Africa, and the Middle East), where non-Indo-Europeans (or Pre-Indo-European) communities were already living for a long time.

Devoto used the term ‘Mediterranean substratum’ to reflect possible linguistic traces of the hypothesised Pre-Indo-European language(s) in (Proto-)Indo-European and in the Indo-European languages. He utilised the notion of ‘Mediterranean’ and extended it also to other regions, such as Anatolia, Iran, and India. Devoto’s theory was popular at the time because it gave a justified relevance to the Celtic element of the Indo-European linguistic layer, by connecting it with the ‘original moment’ of the settlement of the Indo-Europeans in Europe. Despite being influential, and still being used sometimes as a reference to the possible Urheimat (‘homeland’) of the Indo-Europeans, his approach has over time become relatively obsolete.

4.3 Marija Gimbutas’ ‘Kurgan Hypothesis’

Marija Gimbutas developed the so-called ‘Kurgan Hypothesis’, also known as the ‘Kurgan Theory’, ‘Kurgan Model’, or ‘Steppe Theory’, on the origins of the Indo-Europeans. This theory is still one of the mainstream positions on the identification of the possible Urheimat of the Indo-Europeans themselves. She proposed her hypothesis based on very significant archaeological findings, and provided a largely acceptable proposal identifying the possible homeland from where the Indo-Europeans and their languages spread throughout Europe, Eurasia, and parts of Asia. Her position describes the hypothesised arrival of the Indo-Europeans in their new territories at around 4500 to 1000 bce, in different stages, or ‘waves’. Gimbutas postulated this by diachronically ‘following’ various facets of the so-called Kurgan culture she had contributed to unearthing, especially the routes from the east to the west of the burial and funerary traditions (the Turkic term kurgan means ‘burial’) belonging to that culture (Reference GimbutasGimbutas, 1974; Reference Gimbutas1989; Reference Gimbutas1991; Reference Gimbutas1999). These burial sites were accompanied by specific types of pottery and other cultural artefacts. In Gimbutas’ view, what we know as ‘Indo-European’ could, actually, be a (proto-)language, or a set of languages, resulting from the merging of the Indo-Europeans (which the archaeologist identified as Kurgan people) with the Indigenous Old Europeans (Reference GimbutasGimbutas, 1974; Reference Gimbutas1989; Reference Gimbutas1991; Reference Gimbutas1999), who were possibly non-Indo-European (Pre-Indo-European) people.

This hypothetical language, or set of unattested remote languages that we can reconstruct only in the shape of a proto-language, is referred to as ‘Common Indo-European’. This is a sort of synonym of ‘Proto-Indo-European’, and the prefix proto- indicates, as mentioned above, that: (a) it was, theoretically, the language that was commonly spoken by the original Indo-Europeans, who migrated into Europe and all the way to India, before they were separated from each other, with the consequent development of the different Indo-European languages; and (b) we do not have any documents or records of this unattested, hypothetical language, and it can only be reconstructed through the comparative method on the basis of the attested Indo-European languages.

4.4 Hans Krahe’s ‘Old European Hydronymy’

In 1963–4, the German linguist and philologist Hans Krahe conducted a comprehensive analysis of many hydronyms in central and western Europe, from the Baltic to the Iberian Peninsula (Figures 4.2 and 4.3).

Figure 4.2 Krahe’s list of Old European hydronyms for the root *sal-/*salm-

(adapted from Ras67, 2017a)

Figure 4.3 Krahe’s list of Old European hydronyms for the root *al-/*alm-

(adapted from Ras67, 2017b)

These watercourse names constituted what he deemed as the ‘Old European Hydronymy’ (in German, Alteuropäische Hydronymie) (Reference KraheKrahe, 1962; Reference Krahe1964). According to the scholar, these hydronyms contained roots that could belong to a very ancient linguistic layer of Europe. The roots *var-/*ver-, *al-/*alm-, and *sal-/*salm-, which occur repeatedly in the river names he analysed, and the recurring endings in *-a seemed to foreshadow a remote pattern in the prehistoric European toponymic naming processes (Reference KraheKrahe, 1962; Reference VennemannVennemann, 1994). According to Krahe, these hydronymic occurrences suggested that a language (or a few languages belonging to the same linguistic system and possibly sharing basic lexicon and typological features) was prevalent in a large part of Europe in prehistoric times (Reference Mailhammer, Mailhammer, Vennemann and OlsenMailhammer, 2015; Reference Vennemann and NoelVennemann, 2003).

It should be noted that the wording ‘Old European’, according to Krahe, refers to a very old layer of European hydronymic lexemes, which is different from the definition given by Reference GimbutasGimbutas (1989), who would have identified this layer with a non-Indo-European (and/or Pre-Indo-European) Neolithic linguistic context. Reference KraheKrahe (1964) postulated that the language giving rise to the ‘Old European Hydronymy’ cannot be detected anymore. Nonetheless, by comparatively using Indo-European roots and historical-phonetic procedures to analyse the ‘Old European’ hydronyms, Krahe came to the conclusion that these names were not generated from the ‘mother tongue’, or single language, from which all the Indo-European languages came, but belonged to an intermediate Western Indo-European layer that gave rise to most of the West European languages, such as Germanic, Celtic, Italic, and so on.

The linguistic homogeneity observed in the ‘Old European Hydronymy’ led researchers to debate on whether prehistoric Europe was in fact linguistically homogeneous or not. Mitochondrial DNA evidence seems to show that the Indo-Europeans were not very different, genetically, from the populations possibly pre-existing them, whom they subsequently displaced (Reference AnthonyAnthony, 1995; Reference Bentley, Chikhi and PriceBentley, Chikhi, & Price, 2003; Reference Haak, Forster, Bramanti, Matsumua, Brandt, Tänzer, Villems, Renfrew, Gronenborn, Alt and BurgerHaak et al., 2005; Reference Nichols and SpriggsNichols, 1998; Reference SchlerathSchlerath, 1973; Reference Schlerath1981; Reference Schrijver, Carpelan, Parpola and KoskikallioSchrijver, 2001). However, authoritative scholars like Reference TraskTrask (1997, p. 364) opt for a view of ‘Old Europe’ as a land with a diversified set of languages, ‘large and small, some related, some not’. In this context of prehistoric multilingualism, possibly involving the interaction between different (proto-)language families, Basque (the language Trask studied in depth) – or proto-Basque – would represent an almost unique ‘relic’ of that situation and, according to this position, it is plausible to think that languages connected with Basque were more common in Europe (the so-called ‘Old Europe’) before the Indo-European languages replaced them. Basque, therefore, would be a remnant of a diachronic stage preceding Indo-European in Europe, in a moment in which the linguistic situation of the Old Continent was quite fluid. Apparently, no living languages are related to Basque, which is a language that is quite different, in all aspects, from Indo-European languages (despite a long interaction with them, especially with Spanish). Other long-extinct languages, possibly Pre-Indo-European, could have been related to Basque, for instance Aquitanian, but this connection would date back to a time predating the Indo-European settlements in the Old Continent. Therefore, Basque speakers were possibly the original inhabitants of the Iberian Peninsula, way before the place was occupied by the Celts and then by the Romans. The existence of the Basque language would, therefore, be evidence of the diversified linguistic situation of Europe Trask was referring to.

However, the debate on the nature of ‘Old Europe’, on the existence of a language or set of languages that can be defined ‘Pre-Indo-European’, and on Basque itself as evidence of a linguistic layer predating Indo-European in general is still ongoing.

Independently from the exactness of the ethnographic reconstruction, Krahe’s work on ‘Old European Hydronymy’ is still a very valuable reference on the ancient names of European bodies of water, at the linguistic and toponymic level, especially in the context of the analytical investigation of hydronyms. Krahe’s research is still fundamental in the debate on Indo-European origins and on the settlement dynamics of Europe between the so-called ‘last Ice Age’Footnote ⁴ and the Neolithic.

Most scholars of Indo-European agree with Krahe’s analysis according to which the hydronyms he analysed belonged to a remote component of the Indo-European family of languages, and this, with the unavoidable variants, remains the mainstream view. One of the most famous and controversial theories opposing this position is Theo Reference Vennemann and NoelVennemann’s (2003) ‘Vasconic Substrate Hypothesis’. Vennemann’s main argument is that the hydronyms analysed by Krahe belonged not to an ancient layer of Indo-European, but to a Pre-Indo-European language family called, by him, ‘Vasconic’, of which the only surviving language in modern times would be Basque. Venneman’s theory is a unicum in the context of studies on Pre-Indo-European Europe, being divergent from all other approaches so far produced. For this reason and due to several debatable assertions and postulations (i.e., the chronology of the origins and development of Vasconic languages, the notion of a ‘Vasconic language family’ in itself, and the related toponymy), it is extremely original, but also considerably controversial.

4.5 Theo Vennemann’s ‘Vasconic Substratum Theory’

Reference VennemannVennemann (1994) focusses on the role of Afro-Asiatic (or Semitic) and Vasconic (e.g., Basque) languages in the prehistoric development of the languages of Europe. His theory is proposed in a series of works that the researcher wrote over the years and is mainly presented in his famous book Europa Vasconica – Europa Semitica, a collection of 27 essays. Vennemann argues that, after the so-called ‘last Ice Age’, most areas of central and western Europe were inhabited by speakers of a family of languages which he calls ‘Vasconic’, of which the only survivor today, according to him, is Basque. Vasconic speakers formed a very ancient substrate of Indo-European. Vennemann draws his primary evidence for the presence of elements of what he defines ‘Vasconic’ throughout much of Europe from the above-mentioned ‘Old European’ hydronyms collected and analysed by Hans Krahe (Reference KraheKrahe, 1962; Reference Krahe1964). He re-analyses these hydronyms according to a new perspective, and interprets them as ‘Vasconic’. According to the scholar, this toponymic evidence leads to the hypothesis that Basque, generally considered a language isolate with no living related languages, is the only survivor of a larger Pre-Indo-European (‘Vasconic’) language family. He argues that this Pre-Indo-European family of languages once (in very remote times) extended throughout most of Europe, and that it also had some influence on the later Indo-European languages (Reference Vennemann and NoelVennemann, 2003).

Vennemann’s theory ingenerated heated debates, and many scholars discussed it widely. Among them, Reference KitsonPeter Kitson (1996) wrote an important paper challenging the ‘Vasconic Hypothesis’ and developed a thorough re-analysis and discussion of Krahe’s ‘Old European Hydronymy’. He problematised the notions of ‘Pre-Indo-European’ and ‘Old Europe’, suggesting that nearly all of the examples provided by Vennemann were erroneously segmented. He then provided a comprehensive Indo-European interpretation of the ancient European Hydronymy. This not only challenged the theoretical premises and assumptions of the theory proposed by Reference VennemannVennemann (1994; Reference Vennemann and Noel2003), but also provided a more up-to-date reassessment of Krahe’s findings. According to Reference KitsonKitson (1996), all the unusual structures observed in the so-called ‘Old European Hydronymy’ have Indo-European roots. In his effort to explain those place names according to an Indo-European key of interpretation, he also connected words for water bodies of ancient Indo-European languages (e.g., Hittite and Old Iranian) with Krahe’s ‘Old European Hydronymy’, showing, therefore, that the ‘Old European Hydronymy’ in itself could not have been ‘Pre-Indo-European’, because it intrinsically and essentially belonged to the Proto-Indo-European context. His critique of Vennemann is summarised in the following passages.

Kitson concludes his critique by stating that ‘The linguistic material of the alteuropäisch river names is Indo-European, and they must be analysed rationally on that basis’ (Reference KitsonKitson, 1996, p. 113).

The scholar thus disagrees with Vennemann’s hypothesis of the ‘Vasconic’ linguistic layer by affirming that it is morphologically incorrect, before reminding us of the essentially Indo-European nature of Krahe’s ‘Old European Hydronymy’. As mentioned, Reference KitsonKitson (1996) was not alone in his opposition to Vennemann. In more recent publications on the notion of ‘Old European Hydronymy’, Reference BichlmeierHarald Bichlmeier (2012) and Reference UntermannJürgen Untermann (2009), among others, also support the Indo-European origins of European river names. Reference BichlmeierBichlmeier (2012), for example, refutes the comparison of ancient ‘Old European’ place names with elements of modern Basque (Reference BichlmeierBichlmeier, 2012, p. 424). Moreover, Untermann, an authoritative scholar in the field of Paleo-Hispanic languages (and one of Krahe’s former students), provides further linguistic proofs for the Indo-European origins of the ‘Old European Hydronymy’, with unexceptionable examples mainly belonging to the Iberian Peninsula context (Reference UntermannUntermann, 2009).

4.6 Mario Alinei’s ‘Palaeolithic Continuity Theory’, or the ‘Palaeolithic Continuity Paradigm’

Another significant approach to the interpretation of the origins of the languages of (prehistoric) Europe is the so-called ‘Palaeolithic Continuity Theory’ (PCT), or ‘Palaeolithic Continuity Paradigm’ (PCP), known in Italian as Teoria della Continuità. The PCP was developed, in the mid-1990s, by Mario Alinei and his international work-group (Reference AlineiAlinei, 1996; Reference Alinei2000). This theory is based on the postulation that what we call ‘(Proto-)Indo-European’ was already spoken in Europe from the Upper Palaeolithic, with the first inhabitants of the Old Continent belonging to the Homo sapiens species, at least around 40,000 years ago (therefore, with the migration of Homo sapiens in Europe). This would place the possible arrival of what we call ‘the Indo-Europeans’ at a much earlier time than the currently accepted mainstream notion.

The timeline proposed by the PCP, with its consequent linguistic implications, enables a number of problematic etymologies of seemingly morphologically obscure Indo-European common lexical items and place names to be solved and, in a way, explained. An example of this is given by the terms for eel in Indo-European languages, which are difficult to be reconstructed with a traditional approach. The English word eel in itself, indeed, has a relatively obscure etymology, as well as its cognates in other Indo-European languages. These words include, among others, German aal, Old High German and Old Saxon āl, Old Frisian ēl, Dutch aal, Old Icelandic āll, and Danish and Swedish ål. According to the PCP, the Proto-Indo-European *al- (and *alb-) root can effectively explain the etymology of eel. *al-, in Indo-European, is connected with the meaning of ‘feed’, or ‘nourish’, detectable, among others, in the Latin verb alō ʻI feed’, or ‘I nourishʼ, in the Old Irish verb alim ʻI nourishʼ, and in the Old Icelandic form ala ʻ(to) feedʼ. The notion of ‘feed’, or ‘nourish’, in turn was originally closely linked to the concept of ‘water’, and specifically to one of its products, ‘fish’. The eel is a ‘fish’ that lives in the ‘water’, and was part of the diet of prehistoric European people, who by eating it were obtaining a form of nourishment coming from the water. By expanding on this reconstruction, it is therefore possible to hypothesise a close relationship between the Indo-European roots *al- and *alb-, and the possibly equivalent Indo-European stem *albh- ‘water’ discussed later in this chapter. The two roots are probably etymologically and historical-phonetically related, and their connection is made stronger by their semantic link to the notion of ‘water’ and of the ‘nourishment’ which could come from water. The PCP postulates the existence of variants of a single basic root referable to a concept, like ‘water’, and to a toponymic indication of a place situated close to a watercourse or to the sea (represented, in Indo-European, by the etymological alba-type). The theory dismisses the mainstream ‘Kurgan Hypothesis’ proposed by Marija Gimbutas, highlighting the fact that, apparently, there are no genetic and archaeological traces of invasions of different populations in Europe after the so-called ‘last Ice Age’ (therefore, the movements of the Kurgan people would have represented an ‘internal migration’). The PCP supporters claim that this lack of archaeological evidence strengthens the idea of an uninterrupted cultural continuity from the Palaeolithic to Neolithic times.

Despite the active participation of other distinguished linguists, such as Guido Borghi and Xaverio Ballester, in the development of this theory (see also Reference Ballester and BenozzoBallester & Benozzo, 2018), the PCP is not universally accepted and is conversely also rejected by several scholars who do not share the concept of ‘continuity’. The PCP has been criticised on a number of methodological and epistemological issues. For example, at the linguistic level, by pushing back the ‘epoch’ of the presence of the Indo-Europeans in Europe, the PCP assumes that it took a tremendous amount of time for Indo-European to split and differentiate into different branches, and that in Neolithic times the different Indo-European languages had already been consolidated. This point is seen as very problematic by many scholars. Moreover, the etymologies reconstructed through the PCP’s methodology cannot be confirmed generally according to the traditional application of sound laws and historical-phonetic criteria.

4.7 Pre-Languages, Proto-Languages, and Toponymy

The notion of ‘pre-language’ we discussed above is unavoidably problematic. Staying with the European example, no matter what theory is proposed, it is a fact that, before the possible arrival of the Indo-Europeans, Europe was already inhabited, and had been for a very long time, by individuals belonging to the genus Homo. These people spoke their own language (or languages), which are forever lost.

• The literature, generally, refers to ‘Indo-European’, ‘Common Indo-European’, and ‘Proto-Indo-European’ as being the same Indo-European (proto-)language, that is, the unattested, reconstructed (proto-)language theoretically spoken by all the original Indo-European people(s) before its differentiation into second-stage Indo-European proto-languages (e.g., Proto-Germanic, Proto-Italic, Proto-Hellenic) and the attested historical Indo-European languages (e.g., Gothic, Latin, Ancient Greek). This proto-language would include, by definition, etymologies traceable back only to Indo-European itself (not to other language families or proto-languages) and reconstructed up to the level of roots, as, for example, in the Proto-Indo-European Etymological Lexicon by Reference PokornyJulius Pokorny (1969), without taking into consideration possible language contact, in prehistoric times, by the Indo-Europeans during their journey(s) towards Europe.

• We can, at the theoretical level, add to this set another term, ‘Pre-Proto-Indo-European’, which refers to the more ancient variety of Indo-European spoken by the Indo-Europeans before they spread across Europe, indeed, before their possible arrival in the Old Continent, that is, a possible, theoretical stage of Indo-European predating what is called ‘Proto-Indo-European’ or ‘Common Indo-European’, representing a more ancient layer of the development of Indo-European in itself.

• We use, conversely, ‘Pre-Indo-European’ to refer to the (plausibly) non-Indo-European languages spoken in Europe before the arrival of the Indo-Europeans, that is, the possible linguistic stage, in Europe, before the ‘arrival’ of the Indo-Europeans, which was ‘erased’ by the arrival of the Indo-Europeans themselves.

In this section, we will move our discussion to the role of toponymic research in the discussion of pre- and proto-languages. Though limited to their lexical context, etymological reconstructions of toponyms are still important in the postulation of pre-languages. As mentioned above, we have, indeed, very little linguistic evidence (or no evidence at all) of the existence of pre-languages, except probably for some place names that have survived till today and that could carry within themselves traces of those long-lost languages. Ancient toponyms, as we have mentioned thus far, are somehow stable in their basic linguistic morphology over time, and they tend to preserve their roots even when speakers have completely lost the knowledge of their origins and original meanings. This makes them, as we said in Chapter 1, ‘linguistic fossils’ or ‘linguistic relics’ of previous stages of languages spoken by populations which have been superseded by other peoples speaking different languages. For this reason, they can theoretically survive across time, when all other aspects of the languages around them have completely changed. They can, therefore, provide us with some insights into the dynamics and notions of ‘pre-’ and ‘proto-languages’. The etymological reconstruction of place names allows us, therefore, to postulate the existence of roots that originated from pre-languages.

Because pre-languages and proto-languages are by definition different and theoretically unrelated, they cannot be matched phonetically or phonologically. As a result, we have to expect that a toponym coined by speakers of a postulated pre-language should be phonetically and phonologically modified, during the passage between the pre-language context to the new proto-language reality, to be adapted to the new proto-language. Therefore, the pronunciation of the toponyms should be equated to the standard linguistic features of the proto-language itself. However, place names are not common lexical items but ‘special’ linguistic entities that generally do not to follow the ‘normal’ dynamics of diachronic language change. Therefore, it is possible to expect that, despite the unavoidable phonetic and phonological changes, the roots of these toponyms would tend to be preserved.

It is also worth remembering that although toponyms may be relatively stable in their form, their meanings are subject to change and, over time, their original semes, or the original semantics of their roots, may be forgotten by the speakers of a language or by new speakers of another language. Local speakers would try, then, to relate the names that have become ‘obscure’ to more familiar terms in their language and to change linguistic elements of the original toponyms to match them with more familiar lexical items. As a result, they can generate toponymic paretymologies (as discussed in Chapter 3). This is a sort of paradox; while, generally, ancient toponyms are relatively stable and tend to retain their original roots, they are also susceptible to change, especially due to population shifts and contact with other languages. It is up to the toponymist, then, to reconstruct the correct etymology of the place names, uncovering their original morphological forms and meanings.

In the next section, we will present and discuss three specific examples of language change in possible contact contexts, and of how toponyms, while generally retaining their roots, experience shifts in their phonetics, phonology, morphology, semantics, and sometimes orthography.

4.8 The Notion of ‘Pre-Proto-Indo-European’: The Case of the *hₐalbh-/*alb- Toponymic System

Language change, as mentioned in Chapter 2, can happen at a number of linguistic levels: phonological, morphological, lexical, syntactic, and semantic. In Chapter 3 (see also Spigno, in Chapter 7), the case study of Squaneto, in particular, demonstrated a possible phonetic and phonological shift in which speakers could have combined the *s- and *-n- sounds to the root of the place name, akᵂ- ‘water’, with the addition of the suffix -eto (< Latin -etum) indicating the notion of ‘place’, or ‘village’. The phoneme *s- was possibly added as a ‘formant consonant’ and did not belong to the original root of the toponym. The phoneme *-n- was added as a euphonic particle, possibly to avoid the contact of the two vowels ‘a’ and ‘e’ in the name Squaneto, which would have made its pronunciation less easy. The etymology of Squaneto, then, may be confirmed also by the comparison with other place names (and their naming processes) etymologically and geographically related, like Squagiato and Saquana (see also Chapter 6). If we accept this reconstruction, then we can say that these toponyms all belong to a consistent toponymic system.

In this section, we turn our attention to another very interesting (and problematic) toponymic system, which seems to show a large number of places in Europe deriving their names from a single root, *alb- (Table 4.1).

Here, it may be appropriate to briefly recap the definition of the notion of ‘toponymic system’ from Chapter 1. Toponymic systems are sets of place names of any type that belong to a specific geographic area and share the same etymological stem and/or the same original naming process. Places that date back to prehistoric times probably had originally very simple names, with the same name possibly repeated for different places, when that name was indicating ‘something important’ for the local prehistoric speakers. Places in a specific area, therefore, probably shared the same names or had very similar denominations, connected with the primary goods for survival and differentiated only by linguistic features that indicated characteristics of the toponym’s localities, such as a specific geo-morphology and/or very important natural resources. For these reasons, toponymic systems can provide evidence of how early (prehistoric) humans organised their geographical landscape into an ‘ideal, intangible map’, in the absence of written languages and/or records.

The toponymic type alba is the basis for the names of many places in Europe (and also in parts of northern Africa and the Middle East), constituting a large toponymic system. These toponyms are derived, possibly, from the Proto-Indo-European root *alb-/*albh-. Its original meaning is connected with the notion of ‘water’ (Reference Perono CacciafocoPerono Cacciafoco, 2013), with some semantic differences when compared to the Indo-European stem *ak^w-, which simply means ‘water’ (in itself). *ak^w- is, indeed, a sort of ‘pan-Indo-European’ root for ‘water’, and many Indo-European languages have lexical items derived from it. From this root, for example, the words aqua in Latin, acqua in Italian, eau in French, agua in Spanish, and so on, are derived. Conversely, *alb- is an Indo-European root also meaning ‘water’, but its semantics is more specialised, being connected with a ‘quality’ (or ‘colour’) of water, namely its ‘transparency’.

At the theoretical level, the Proto-Indo-European root *alb- could be interpreted as deriving from a hypothetical Pre-Proto-Indo-European stem *hₐalbh-, also meaning ‘water’. While the notion of ‘Pre-Proto-Indo-European’ (as defined by us above) is disputed, in the field of historical linguistics, it could be considered plausible in this reconstruction.

Reference Perono CacciafocoPerono Cacciafoco (2013) uses the notion of ‘Pre-Proto-Indo-European’ as a terminological postulate for a diachronic analysis of the *alb- toponymic and hydronymic system. Elements of this system can also be found along what could be possible migration routes (including territories located outside of the Old Continent) taken by the original Indo-Europeans in their migrations from outside Europe. Since the notion of ‘Pre-Proto-Indo-European’ is only theoretical, this is indeed an experimental reconstruction. In this context, this concept can be used to postulate a stage of Indo-European before language contact with other prehistoric (proto-)languages and/or language families. Within this framework, Perono Cacciafoco postulates the existence of a Pre-Proto-Indo-European root, *hₐalbh-, etymologically more ancient than *alb-/*albh-, which would have generated the Proto-Indo-European stem (*alb-/*albh-). This could have happened during the Indo-Europeans’ movements towards their historical venues and before their possible arrival in their historical territories. The root, therefore, would have developed due to language contact with other (proto-)languages or language families during the Indo-European migrations. The stem, at the same time, would have been at the origins of the development of the above-mentioned toponymic system, documented also by the related roots and proto-forms *olbh- and its own variants *olb- and *orb- (Reference Gasca Queirazza, Marcato, Pellegrini, Petracco Sicardi, Rossebastiano and PapaGasca Queirazza et al., 1999; Reference PellegriniPellegrini, 2008), *albhā- (a late Indo-European form meaning ‘city located on the water’, or ‘town located on the water’ and later simply ‘city’, or ‘town’), and *albho- ‘white’ (> Latin albus). These will be discussed in more detail below.

Actually, stems and lexemes could have been borrowed from non-Indo-European languages by the Indo-Europeans and reused and re-functionalised in their linguistic system during their journeys before their possible arrival in Europe and India. In particular, the stem *hₐalbh-, despite being in its nature Indo-European, could be compatible with non-Indo-European lexemes like the Sumerian word ḫalbia, from which the Akkadian ḫalpium would have derived, with both lexemes meaning ʻspring, well, water mass, water holeʼ. The geographical point of contact between non-Indo-European linguistic contexts (Semitic, in the case of Akkadian) and Indo-European with the development of the root *hₐalbh- could have been Anatolia, and could be linguistically (at the level of attested lexemes) represented by the Hittite (Indo-European) term alpa-s, meaning ‘cloud’ (and, by extension, ‘rainwater’). According to this theoretical reconstruction, therefore, the Pre-Proto-Indo-European root *hₐalbh- would have been adopted by the Indo-European speakers in their linguistic system, producing the normalised form *alb- (~ *albh-).

The postulated Pre-Proto-Indo-European root *hₐalbh- and the Indo-European stems *alb- and *albh- are indeed equivalent to each other and originally all indicated the notion of ‘water’. Moreover, as seen from the list of place names in the toponymic system represented in Table 4.1, they were very productive in toponymy and, as names of places, used to indicate the notions of ‘place located on the water’ and ‘place bordering a water body’. That is, they are metonyms, names for inhabited places with a primary etymological sense of ‘water’.

To account for the variants of *alb- that begin with /o/, for example, Olbia, the oldest colony of Miletus, on the Black Sea, it is possible to hypothesise that the root underwent a sort of vocalic ablaut (apophony) of the initial vowel /a/ into /o/. The *olb- variant is found in many place names in Europe, such as the other multiple occurrences of the olbia-type outlined in Table 4.1. There are also place names belonging to this type, such as Orba and Orbicella, in which the phoneme /l/ changed into /r/ in postvocalic contexts and before bilabial consonants. Such a linguistic phenomenon is relatively common in the area of ancient Liguria, Italy, which is now located between the lower Piedmont and the contemporary Liguria region, where a number of these toponyms are found.

The root *olb- is, therefore, equivalent to the Ligurian stem *orb-, as witnessed, for example, in the name of the Orba River. The first known attestation of the hydronym dates back to 1137, with the form Urba (VV.AA., 1899–, CXIII, 38, 53). In 1176, it was indicated by the denomination Urbae (VV.AA., 1899–, XXIX, 74, 94). The common etymological explanation of this hydronym is that it derives from the Latin word urbs, meaning ‘city’, or ‘town’. However, as explained in this volume, many reconstructions of (peripheral) place names have been carried out without taking into consideration historical-linguistic factors and only by linking toponyms to a contemporary stage of the local language or to its immediately preceding stage. The name of the river is clearly connected with the notion of ‘water’ and, therefore, the medieval form Urba is a paretymological variant, produced by the speakers to connect the hydronym with a ‘safe’ (but wrong) referent, the Latin word urbs. Nonetheless, the river name clearly derives from the root *alb-, meaning ‘water’, in its variant *olb-/*orb-. Another hydronym that shows the link between the variants *olb- and *orb-, and the stem *alb-, is Orbicella, the denomination of a stream which flows in the territory of the village of Olbicella (which has the same name as the hydronym, just with the variant *olb-) and is an affluent of the Orba River. The hydronym itself is further linguistic evidence proving the equivalence of the roots *olb- and *orb- (< *alb-), and illustrates the coherence of the sound change from /l/ to /r/. A similar reasoning can be used to account for the name of the Urbe municipality, located in the valley of the Orba River. The place name is paretymologically connected, by the local speakers, with the Latin word urbs ‘city’. This was probably due to the perception that the name was associated with the more ‘prestigious’ Urbs, with a capital ‘u’, that is, Rome itself, therefore meaning ‘the City’. The denomination instead etymologically coincides with the hydronym Orba and derives from the same root (therefore, Urbe = Orba < *olb-/*orb- < *alb-) and not from Latin urbs (Reference Gasca Queirazza, Marcato, Pellegrini, Petracco Sicardi, Rossebastiano and PapaGasca Queirazza et al., 1999).

The stratification of the meaning of the root *alb- can be seen happening in late Indo-European, with place names not directly associated with ‘water’ but, in fact, connected with the Indo-European proto-forms *albo(-), *albio(-), and *alba(-), with the meaning of ‘city’, or ‘town’. This can be seen, for example, in the names of three towns from northwest Italy, Albera Ligure (where the Albirola stream, another hydronym derived from *alb-, flows), Albenga (in Latin Album Ingaunum), and Ventimiglia (in Latin Album Intimilium). The alba/album component, in these place names, comes from the root *alb-, but is not connected anymore (primarily) with the notion of ‘water’, meaning rather ‘city’, or ‘town’, or ‘inhabited centre’. These toponyms, in association with human settlements, therefore underwent the semantic transition of the seme of the root, from the meaning of ‘water’ to ‘place located in the closeness of water’ (because these inhabited centres were and are located, for self-evident reasons, in proximity of water) and then to a more generic meaning of ‘place’, ‘city’, or ‘town’. This phenomenon can be observed in a number of toponyms across Europe belonging to the *alb- toponymic system. The proximity to waterways and water bodies has always been fundamental to the birth and development of villages, towns, and civilisations in general.

In the context of the semantic shift connected with the *alb- toponymic system, Reference Perono CacciafocoPerono Cacciafoco (2013) notes that another phenomenon of meaning change can be highlighted through the analysis of the late-Indo-European adjective *albho(-) ‘light’, or ‘white’. This proto-form also derives from the root *alb- (~ *albh-) ʻwaterʼ, but its meaning indicates a colour and no longer the primary resource (water). However, its origin has to be still connected with a characteristic of the water, specifically its ‘colour’ or, better, its qualities of ‘transparency’ and ‘purity’ (i.e., ‘clear water’, ‘white water’). The root, therefore, underwent another semantic change in the development of this proto-form, from ‘water’ to ‘white’, in the late Indo-European proto-form *albho(-) for ‘light/clear/white water’, in the equivalent proto-form *albha(-). This reconstruction, involving the generation of place names and common lexical items from the root *alb-, suggests a diachronic stratification of different meanings for the same root. Below, we outline the semantic transition that led to this, in the passage from ancient and remote phases of Indo-European to late Indo-European.

Another point worth noting is that place names derived from the root *alb- are also part of a series of toponyms, from different Indo-European contexts, that relate to other water-roots (different from *alb-), which are radicals with the original seme of ‘water’. These other roots are semantically specialised according to specific meanings, such as *war- ‘water’, ‘flowing water’, or ‘river’, or ‘rain’, *pal- ‘stagnant water’, ‘puddle’, or ‘backwater’, *mar- ‘lagoon’, or ‘sea’. Some of them indicate characteristics or qualities of water or of water currents, such as, for example, *tar- ‘strong’, or ‘penetrating’, and *ais- ‘fast’ (Reference VillarVillar, 1996, p. 117), like in, for example, Tarleton, in northwest England, and Aix-en-Provence (Ais, in Mistralian, Aquae Sextiae, in Latin) and Aix-les-Bains (Aquae Gratianae, in Latin), in France.

All these roots, in general, are at the origins of toponymic systems, naturally connected with hydronymy. If we analyse, for example, the *war- root, we can find toponymic examples in two hamlets of the Urbe municipality called Vara Inferiore and Vara Superiore, both associated with this root (Vara means ‘flowing water’, and comes from the root *war-, while Inferiore means ‘lower’ and Superiore means ‘upper’). The root is also the basis for a number of hydronyms, such as the Vara River, the longest watercourse flowing in Liguria, Italy, and the Varàita River, flowing in the province of Cuneo, Piedmont, in northwest Italy. As is evident from this little sample of the toponymic system connected with *war-, the notion of ‘water’ was always present as an important naming component of place names from ancient times, always highlighting the importance of this primary good that makes life possible.

An interesting case suggesting the persistence of the meaning in the naming process is represented by Acquabianca, the name of a hamlet belonging to the above-mentioned Urbe municipality. The toponym, Acquabianca, is a compound which literally means ‘white (bianca) water (acqua)’, and comes from two forms independent from the root *alb-. Acqua is the Italian word for ‘water’ and derives from the above-mentioned Indo-European root generically meaning ‘water’, *ak^w-, which developed into Latin aqua, which consequently generated the Italian acqua. On the other hand, bianca (bianc-o, ‘white’, masculine singular; in Italian, the ending -a marks the feminine singular form) comes from Old High German blanch ‘white’, which then became the Modern High German form blank ‘bright’ or ‘shiny’. The Germanic word was introduced in the Latin/Italian lexicon in the High Middle Ages. Despite the name Acquabianca not being connected etymologically with *alb- and its toponymic system, its meaning is linked to the ‘colour’ of water, the ‘white’/‘transparent’ colour semantically indicated by the *alb- root. Independently from the historical morphology of the name, the cognitive and semantic components of its naming process are the same as the other place names derived from *alb-. That is, even as they are different at the morphological level, they are all connected by the notion of the ‘white’ (‘transparent’) colour of the water. This would indicate, therefore, a semantically equivalent naming process produced by the speakers at the cognitive level by using different morphological units.

In summary, we have reviewed a number of toponyms that are all connected with each other by the Proto-Indo-European root *alb- (~ *albh-) ‘water’ (with the variants *olbh-, *olb-, and *orb-). The *alb- toponymic system consists of place names (town names, village names, hamlet names, and hydronyms) that are characterised by very similar or even the same names. Our discussion highlighted not only the existence of this toponymic system, but also the semantic shifts involving the root *alb- in two directions, over time. One change follows the sequence:

while the other change, instead, follows the sequence:

This case study provides us with an excellent example which highlights the importance of waterways in the development of prehistoric settlements and which reconstructs a very ancient naming strategy based on the relationship between place names and the natural features of their locations. At the same time, it also shows the effects of language change over toponyms and lexemes belonging to the common vocabulary of languages, particularly the strength of semantic change.

4.9 The Toponymic Persistence of Prehistoric Stems: The Case of the *kar- Root

Reference Perono CacciafocoPerono Cacciafoco (2015a) provides a consistent discussion on what can be called the ‘toponymic persistence’ of linguistic roots and stems from very remote times. He does so through the study of the Proto-Indo-European root *kar- (with its possible variant *kal-), meaning ‘stone’, or ‘rock’, found in prehistoric European toponyms. Indeed, numerous hydronyms and toponyms connected with the *kar- root exist in Europe. These include, among many others, Harund (Scandinavia), Carron, Cart Water, and Cary (Great Britain), Carad (Ireland), Harste (Germany), Chiers (Belgium), Charentonne, Cher, Charante (France), Carusa (Italy), Carranzo, Carranza, Carisa, and Carantó (Spain). These place names, according to Antonio Tovar, are all believed to be derived from the proto-form *kar(r)a (< *kar-) ‘stone’, or ‘rock’ (Reference TovarTovar, 1977; Reference Tovar1982). Francisco Villar adds another famous denomination to this possible toponymic system, that is, Carrara (central Italy, Tuscany, in the province of Massa-Carrara), the name of a town well known for its celebrated marble (Reference VillarVillar, 1997, p. 126). Additionally, Reference CoatesCoates (2020) recently explored the possible link of a number of Croatian toponyms to the *kar-/*kal- root. All these place names (and many others) are associated with the root *kar- and with the notion of an important primary resource, ‘stone’ (or ‘rock’). While these names are Indo-European in their reconstructable forms and origins, some of them can be hypothesised to date back to Pre-Indo-European times, because they indicate places that were inhabited in extremely remote times.

As we saw with the *alb- root in the previous section, it is a theoretical possibility that some Indo-European toponymic roots could belong to a Pre-Indo-European linguistic layer of Europe. These stems, involved in the making of toponyms, over time and following the possible arrival of Indo-European people in the Old Continent, would have been incorporated into the Indo-European linguistic system and therefore phonetically, phonologically, morphologically, and semantically reused and re-functionalised by the Indo-Europeans. One such root is *kar- (and see the discussion of *borm-, Section 4.10). To avoid any kind of misunderstanding, all the *kar- place names mentioned above are unmistakably phonetically Indo-European. However, they are so ancient and so widespread all over Europe that they allow us, as mentioned, to hypothesise that they existed before Proto-Indo-European times, that is, before the possible arrival in Europe of the Indo-Europeans, and therefore to postulate that these names could have been originally Pre-Indo-European.

The root *kar- is productive not only at the toponymic level, but also at the level of general lexicon. Examples include the Indo-European proto-form *kartu(-), which is at the origin of the Ancient Greek adjective κρατύς (kratýs) ‘strong’, and of the verb καρτύνω (kartýnô), or κρατύνω (kratýnô), ‘to strengthen’, of Gothic hardus ‘hard’, and of Latin ardŭus ‘hard’. From these instances, we can see that *kar-, along with its derived forms, has the complementary meaning of ‘(being) hard’, which is a typical property of the stone, that is, the hardness of the stone, which is linguistically indicated by *kar- (Reference BenvenisteBenveniste, 2001).

The *kar- root has a presumably LigurianFootnote ⁵ variant, *kal-, which is attested, for example, in the Italian word calanco. Scholars Giacomo Devoto and Gian Carlo Oli define calanco as a ‘narrow and deep erosion’s furrow with many ramifications, limited by thin ridges, generally devoid of vegetation; it is a phenomenon from predominantly clay soils, produced by runoff and/or washing out waters’ (Reference Devoto and OliDevoto & Oli, 1975, p. 406). Calanco, in its plural form calanchi, refers specifically to sedimentary rocks and soils that have been eroded by wind and water. The calanchi are widespread, among other places, in the Appennino Emiliano-Romagnolo (a mountain range in north-central Italy) (Reference Devoto and OliDevoto & Oli, 1975, p. 406), as well as in much of the Liguria region and in the lower quadrant of Piedmont, where many toponyms derived from *kal- can be recorded (Reference Perono CacciafocoPerono Cacciafoco, 2015a). Thus, it is generally accepted that the term calanchi is directly related to these particular hydro-geo-morphological features prevalent in lower Piedmont, Liguria, and Emilia Romagna (northern and north-central Italy). Devoto and Oli believed that the word calanco derives from a possibly Pre-Indo-European (or ‘Mediterranean’, as discussed above) word cala (from *kal- or *kar-) (Reference Devoto and OliDevoto & Oli, 1975, p. 404), with the Pre-Indo-European (then Ligurian) suffix -anco. Another scholar, Tristano Bolelli, writes about the possibly Pre-Indo-European origins of the term cala (Reference BolelliBolelli, 1995, p. 74). These elements depict the ‘oscillations’ in the interpretations of the Pre-Indo-European versus Proto-Indo-European nature of the root(s) *kar-/*kal-. However, as Reference Perono CacciafocoPerono Cacciafoco (2015a, p. 38) notes, ‘nothing is known (with the exception of the hypothesis of their existence) about the languages possibly spoken in the whole Neolithic Europe (and also in more ancient and remote times, starting, at least, from the Upper Palaeolithic) and about their mutual relationships’. Instead, what is known about these languages comes from possible loan-words or loan-roots, like *kar-/*kal- in Indo-European, if we think that *kar- and *kal- came from a Pre-Indo-European substrate.

The connection of *kal- with common lexemes in Indo-European languages indicating different kinds of stones and their properties is further evidence of the association of *kar- with *kal- and of their semantic and cognitive link to the notion of ‘stone’. This is highlighted, among others, by the Latin forms calx ‘lime’ and calcārĭus ‘calcareous’, the Ancient Greek word χάλιξ (chálix) ‘pebble, stone dissolving in water, gravel, lime, limestone’, and the Italian term calcare ‘limestone’ (Reference Devoto and OliDevoto & Oli, 1975, p. 408; Reference BolelliBolelli, 1995, p. 75). According to Reference Perono CacciafocoPerono Cacciafoco (2015a, p. 48), ‘the root *kar-/*kal- seems to be, therefore, a single root expressed by two equivalent and homologous variants’. Through his research on this issue, Perono Cacciafoco was able to add at least nine other toponyms connected with the *kar-/*kal- root to a list of place names constituting the related toponymic system. Among them, the scholar reconstructed the etymology of two Ligurian toponyms, Carcare and Cairo Montenotte, which are the names of places surrounded by the massive presence of calanchi, the hydro-geological phenomena outlined above that have characterised (among other localities) Liguria from prehistoric times.

Carcare and Cairo Montenotte are located in Liguria, northwest Italy (Savona province). They are two towns belonging to the so-called Bormida Valley, which has its name from the Bormida River (the name of this river and the significance of its root, *borm-, will be covered in the next section).

Carcare is located at an important road junction along the route connecting southern Piedmont with the Ligurian seacoast, in the Ligurian hinterland. Historical sources show that one of the oldest versions of Carcare’s name was Carcaris. This was a Late Latin plural ablative attested in a document of the year 1111. Another document also mentions Carcaris in 1179 (Reference Gasca Queirazza, Marcato, Pellegrini, Petracco Sicardi, Rossebastiano and PapaGasca Queirazza et al., 1999). A prehistoric settlement in Carcare’s territory is attested from the Neolithic. Carcare, therefore, was already inhabited from remote ages, and it is possible to hypothesise human presence in its area probably from the Upper Palaeolithic (Reference BiagiBiagi, 1980; Reference Guidi and PipernoGuidi & Piperno, 2005). The toponym could be explained in connection with the hydro-geo-morphology of its territory, characterised by the presence of calanchi and rocky hills, and by hypothesising the reduplication of the *kar- root from which the place name derives.

What makes the toponym Carcare so important is that the phenomenon of reduplication is not common in Indo-European, and therefore is not morphologically productive (Reference Perono CacciafocoPerono Cacciafoco, 2015a, p. 40). Nonetheless, because the reduplication in Carcare’s toponymic form is undeniable, it is possible to postulate that this phenomenon could come from a conceivable Pre-Indo-European morphological pattern. Therefore, the root *kar-, which is undeniably Indo-European, could hypothetically have Pre-Indo-European origins, identifiable through the phenomenon of reduplication detectable in Carcare’s name. That is, a morphological feature not common in Indo-European. This, of course, happens not at the level of etymology, but at the level of ‘morphological treatment’ of the stem. The location where Carcare is ubicated is, as mentioned, surrounded by ‘stony’ hills characterised by the widespread presence of calanchi, which could have been the original motivation for its name. Perhaps the possible reduplication of the root of the toponym sought to emphasise the large amount of calanchi that surrounded Carcare and, if it were so, this would generate a non-Indo-European morphological phenomenon. Indeed, the fact that the reduplication of a root is not a common feature of the Indo-European historical morphology suggests that Carcare could have been part of its geographical and toponymic landscape already in Pre-Indo-European times. Then, after the Indo-Europeans ‘arrived’ and settled in Liguria, they assimilated the root into their linguistic system and, eventually, the radical became a Proto-Indo-European root. However, the stem still retained some of its original morphological properties (e.g., the reduplication) even when it was adapted into the linguistic system of the new speakers.

Cairo Montenotte is a town located near Carcare, in Liguria, in the province of Savona. The toponym has been continuously attested since the Middle Ages (Reference Perono CacciafocoPerono Cacciafoco, 2015a, p. 42). Planimetries of the village from the archives of the Diocesi di Acqui Terme ‘Roman Catholic Diocese of Acqui Terme’ and of the Diocesi di Savona-Noli ‘Diocese of Savona-Noli’ drawn in the tenth century show a medieval town centre installed into the rectangular construction plan typical of feudal planned villages. Meanwhile, Montenotte is the name of a small hamlet belonging to the municipality, famous for a Napoleonic battle that was fought there in 1796. The first attested written records of Cairo Montenotte are from the year 967, as Carium, derived from the root *kar- and therefore linked to the notion of ‘stone’, ‘cliff’, or ‘rock’ (Reference Gasca Queirazza, Marcato, Pellegrini, Petracco Sicardi, Rossebastiano and PapaGasca Queirazza et al., 1999, p. 114), and as Carius (Reference Ravera and TascaRavera, Tasca, & V., 1997, p. 70), with the same meaning. In its original form, Carium/Carius, the place name Cairo (Montenotte) derives from *car(i)- < *kar- (with *kar- being the original stem). The toponym, in the form Cario, is later attested in the year 991, in the Charta di fondazione e donazione dell’Abbazia di San Quintino in Spigno (the ‘Founding and Donation Charta of the Saint Quentin’s Abbey in Spigno Monferrato’) (Reference BosioBosio, 1972). This is the Charta outlining the donation of a large amount of land for the founding of the Abbazia di San Quintino ‘Saint Quentin’s Abbey’, located in Spigno Monferrato (Reference BosioBosio, 1972, pp. 140–2). Both the abbey and Spigno Monferrato are not far from Cairo Montenotte, on the route from lower Piedmont to Savona. The tracing of Cairo Montenotte’s etymology to the *kar-/*kal- root is supported by the hydro-geo-morphology of its territory, characterised by the presence of calanchi, and by the existence of a ‘parallel oronym’ (at the etymological level, even if quite distant geographically), Monte Cairo ‘Cairo Mountain’, a calcareous mount that rises north of the town of Cassino in the Lazio region, central Italy, which shares with Cairo Montenotte the same etymology derived from the root *kar-. Moreover, the geological structure of Monte Cairo is composed of limestone rock, which semantically matches with some of the Ancient Greek, Latin, and Italian words we have listed above, derived from the *kar-/*kal- root. The oronym Monte Cairo, hence, is indirect evidence of the connection of the root *kar-/*kal- with the toponym Cairo (Montenotte).

Aside from Carcare and Cairo Montenotte, many other place names in Italy (particularly from Piedmont, Liguria, and Tuscany) and in Europe in general (as listed above) are derived from the same *kar-/*kal- root. They possibly form a large toponymic system. Such places include, among many others, Carretto, hamlet of Cairo Montenotte, the ancient village of Calizzano, in Liguria, in the province of Savona, Carezzano and Caranzano, in southern Piedmont, Calasca, in northern Piedmont, Charance, an Alpine locality in the Gap municipality, France, and Calci and Carrara, in Tuscany. Of these toponyms, Charance might appear at first sight atypical in the *kar-/*kal- system. Yet, Reference Perono CacciafocoPerono Cacciafoco (2015a, p. 47) explains that the toponym also derives from the root *kar-/*kal-. Specifically, he notes, the original stem for Charance is an Alpine root *cal- (< *kar-/*kal-) and can be traced back to ancient Ligurian word kalanco or kalanca ‘steep stony descent that serves as a channel for avalanches’, or calanco. The meaning of Charance can be, therefore, ‘site of the avalanche of stones’, or ‘place of the stony landslide’, and hence is always connected with the notions of ‘stone’ and ‘rock’, represented by the root *kar-/*kal-.

In this historical reconstruction of two place names in the Ligurian hinterland, Carcare and Cairo Montenotte, we see how toponymy can help us to hypothesise the possible existence of pre-languages. We first analysed the possibility of an interesting reduplication process of the *kar-/*kal- root, extremely uncommon in Indo-European, in the place name Carcare. This morphological structure can, hence, be regarded as a ‘linguistic fossil’ showing that the root could have preserved some of its Pre-Indo-European features even in Indo-European times. We then saw how the root *kar-/*kal- shows its persistence and pervasiveness, in Indo-European context, in the toponym Cairo (Montenotte) and in a large number of Italian and European place names. Indeed, the notions of ‘stone’ and ‘rock’, and the relationship with hydro-geo-morphological features like the calanchi that are widespread in Liguria and elsewhere, are almost always at the origins of the place names derived from the root *kar-/*kal- discussed in this section.

4.10 From Pre-Language to Proto-Language: The Case of the *borm- Root

We just saw, in our analysis of the *kar-/*kal- root, that Indo-European stems could have theoretically been Pre-Indo-European in their origins, and then assimilated over time by Indo-European speakers to become properly Indo-European. The following example is inherent in another place name which could provide some clues on the possible existence of a pre-language (Pre-Indo-European) and of linguistic contact between speakers of that pre-language (Pre-Indo-European) and speakers of a proto-language (Indo-European), when the former language was replaced by the latter.

The hydronym under study in this section is Bormida (see Figure 4.4), the name of an Italian river flowing in northwest Italy, between the regions of Liguria and Piedmont. The denomination is derived from the root *borm- (> *bormo-), meaning ‘warm’, or ‘hot’, or ‘warm water’, or ‘hot water’, and could specifically refer to the waters of the river being ‘warm’, or ‘hot’. Although the waters of the Bormida are not hot, the course of the river is dotted by hot springs, which could have been at the origins of the meaning ‘hot waters’. The stem *borm- (> *bormo-) could possibly also be considered Pre-Indo-European in its origins, or at least ‘more ancient’ than Indo-European (and therefore non-Indo-European), as we will argue below. At the same time, the root also generates an extensive meaning connected with the notion of ‘hot spring(s)’. The Bormida River has two main branches, La Bormida di Spigno and La Bormida di Millesimo, which join near the village of Bistagno, but in the territory under the jurisdiction of the inhabited centre of Sessame.

Figure 4.4 Location of the Bormida River

(adapted from Reference CeragioliCeragioli, 2014; d-maps, 2021)

The *borm- root is attested not only in the hydronym, but also in a number of other place names, all of which are associated with the notion of ‘hot water’, or ‘warm water’. One of them is Bormio, the name of a town located in Lombardy, northwest Italy, in the province of Sondrio. Bormio has been renowned from ancient times for its thermal springs.

Place names associated with the *borm- root include, among others, Burmu, a hamlet of Pigna, Liguria, northwest Italy, in the province of Imperia. Giulia Petracco Sicardi studied Burmu, meaning ‘big spring’ (Reference Petracco SicardiPetracco Sicardi, 1962, p. 63). As stated above, the root *borm- could have an extended meaning connected with the notion of ‘spring’. Meanwhile, Reference Perono CacciafocoPerono Cacciafoco (2015b) lists two place names, Bormida, a village located in Liguria, in the province of Savona (which gets its name directly from the hydronym, being ubicated along the river), and Worms (a famous and ancient city located in the Rhineland-Palatinate area, Germany, on the Upper Rhine), both of which derive from *borm-. It is important to note that the original (Romanised) attested Celtic name of Worms was Borbetomagus, or Bormetomagus, meaning ‘town located in a territory rich in (warm) waters’. The centre, whose official Roman name was Augusta Vangionum, was then called Vormatia, later Wormatia or Wormazia, and the latter name has been in use since the sixth century. From the sixteenth century, the name appeared in the form Wormbs, which leads to the current denomination being attested from possibly at least the nineteenth century.

The hydronym Bormida appears in written sources as Burmia, in line 12 of the above-mentioned Founding and Donation Charta of the Saint Quentin’s Abbey in Spigno Monferrato, drawn up on 4 May 991 (Reference BosioBosio, 1972, pp. 18, 30, 138–40). Burmia is also in an 1137 document (VV.AA., 1899–), and is transcribed as Burmea in 1182 (Reference OlivieriOlivieri, 1965, p. 97). One of the most ancient attestations of the toponym Bormio is Burmis, from 822 (Reference Gasca Queirazza, Marcato, Pellegrini, Petracco Sicardi, Rossebastiano and PapaGasca Queirazza et al., 1999). Going back further in time, Cassiodorus, in his work Variae, which was published in 538, mentions two toponyms, Aquae Bormiae and Aqu(a)e Bormidae (Reference CassiodorusCassiodorus, 1583). Reference Perono CacciafocoPerono Cacciafoco (2015b, p. 347) writes that these two places are difficult to locate and identify, but that both are related to the *borm- root and, hence, connected with either the place name Bormio and/or the hydronym Bormida. Despite the names being geographically difficult to pinpoint, Cassiodorus links Aquae Bormiae to a hot-water spring that can cure gout. Given that Bormio has very renowned thermal springs, this could be a significant clue for the possible association of Aquae Bormiae with Bormio. The name can also be referring to the Bormida River, since the watercourse flows through the city of Acqui Terme (Piedmont, northwest Italy, in the province of Alessandria), a small town renowned already in Roman times for its thermal waters (and for its hot, sulphurous spring called La Bollente ‘The Boiling’). The meaning of the *borm- root is connected with the notions of ‘warm’, or ‘hot’ (in Proto-Indo-European *g^whermó-/*g^whormo-) and thus could be extended to mean ‘warm water(s)’, or ‘hot water(s)’, because of the hot springs which surround the river in a large portion of its course. Getting back to Aquae Bormiae and Aqu(a)e Bormida, these names could therefore definitely be explained as ‘warm waters’, or ‘hot waters’. Agreeing with the fact that Aquae Bormiae and Bormida stem from *borm-, Giacomo Devoto translated Bormida into ‘the hot waters river’, linking his interpretation of *borm- to Aquae Bormiae and the Latin term formus, meaning ‘warm’ (Reference DevotoDevoto, 1962, p. 199). The Bormida River could also be linked to the names of the Celtic and Lusitanian gods Bormō and Bormānus, both associated with healing and (hot) springs. Variants of these names, such as Borūs, Borvō, Bormō, and Bormānus, are based on the *borm- (> *bormo-) root and on a connected stem *boru-. The stem *boru- could derive, in turn, from the *borm- root, and is considered a variant of the Proto-Celtic root *beru- ‘to boil’ (Reference PokornyPokorny, 1969).

The *borm- (> *bormo-) root could also be an equivalent of the Proto-Indo-European stem *bhreue- (and its variants *bher-, or *bhor-) (Reference PokornyPokorny, 1969, pp. 171–2), which means ‘to bubble’, or ‘to boil’, and is connected with words like Ancient Greek φρέαρ (phréār) ‘well’, or ‘spring’, Latin fervēre ‘to boil’, or ‘to foam’, and Gaulish boruo ‘bubbling spring’, or ‘hot spring’. The root *bhreue- (and its variants *bher-, or *bhor-) is also attested in some toponyms in the *borm- toponymic system. Olivieri links the hydronym Borbèra, the name of a stream flowing in the homonymous Borbèra Valley (Piedmont, northwest Italy, in the province of Alessandria) to the root *borb- (which is an equivalent of *borm-, like in Bormetomagus/Borbetomagus), ‘noise of the waters’. This is an onomatopoeic expression connected with the typical gurgling of a (hot) water spring or watercourse, and in turn can be linked to *bhreue- (and its variants *bher-, or *bhor-) ‘to bubble’, or ‘to boil’, a process that produces the typical noise and, hence, the onomatopoeic expression (Reference OlivieriOlivieri, 1965, p. 95).

The *borm- root (as well as the standard equivalent Indo-European stems *g^whermó-/*g^whormo-) also has a role in the ‘making’ of lexical items, in Indo-European languages, connected with the notions of ‘hot’ and ‘warm’, such as the already-mentioned Latin term formus ‘warm’, which can be traced back to the Proto-Indo-European proto-form *g^whermos(-). In this proto-form, the voiced aspirate labiovelar consonant /g^wh/ replaces the voiced bilabial /b/. Indeed, the root *borm- (> *bormo-) and the stems *g^whermó-/*g^whormo- are both Indo-European and are almost equivalent in their meaning and in their morphology. The difference in the root consonant (/b/ vs. /g^wh/) nonetheless seems quite relevant, as well as the slight semantic variation (‘hot water’/‘warm water’ vs. ‘hot’/‘warm’) and the fact that the *borm- root seems more productive at the level of toponymy. Indeed, it is involved mainly in the ‘making’ of place names, while the *g^whermó-/*g^whormo- stems are connected with terms from the general Indo-European lexicon. These differences, therefore, allow us to hypothesise that the *borm- (> *bormo-) root is connected with the specific notion of ‘water’ (‘warm water’, or ‘hot water’, or, extensively, ‘hot spring’) and to a very ancient layer of the toponymy and hydronymy of Europe. Hence, we can postulate that it has Pre-Indo-European origins, and it would have been transferred into the Proto-Indo-European linguistic system through a process of reuse and re-functionalisation operated by Indo-European speakers once they settled in their historical territories in Europe. The root would have then been ‘equated’ by the typically Proto-Indo-European forms *g^whermó- and *g^whormo-, in a more generic and extensive meaning of ‘warm’, or ‘hot’. The (Proto-)Indo-European roots *g^whermó-/*g^whormo-, as well as the possibly more ancient (and originally Pre-Indo-European?) *bhreue- (and its variants *bher-/*bhor-), could have been later linked to Bormida and Bormio (and to their original root, *borm- (> *bormo-)), possibly also through Lepontic (or Celtic-Ligurian), an epigraphic Indo-European language and variety of Celtic. Lepontic is an ancient Alpine language that was spoken in parts of Rhaetia and Cisalpine Gaul (what is now northern Italy) between 550 and 100 bce and that could have inherited the two analogous stems from the possibly Pre-Indo-European root *borm- (> *bormo-) (Reference Gasca Queirazza, Marcato, Pellegrini, Petracco Sicardi, Rossebastiano and PapaGasca Queirazza et al., 1999, p. 92; Reference Pellegrini and CampanilePellegrini, 1981, p. 38).

In conclusion, the stem *borm- (> *bormo-) could have theoretically been widespread in Europe before the possible arrival of the Indo-Europeans. The phonetic treatment of the root consonant /b/ of *borm- seems not to be immediately attributable to the (Proto-)Indo-European historical phonetics and the root would have been incorporated, over time, by the Indo-Europeans into their linguistic system. The stem was then reused and re-functionalised, becoming ‘Indo-Europeanised’ through the stem *g^whermó-/*g^whormo-, meaning, generically, ‘warm’, or ‘hot’.

4.11 Summary

In this chapter, we have explored the differences between the notions of ‘pre-language’ and ‘proto-language’ and, by using examples from the Pre-Indo-European and Proto-Indo-European contexts, we have shown how the study of toponymy can be very relevant in reconstructing forms that can provide us with some important clues on the existence of pre-languages in prehistoric times.

The three toponymic systems analysed, the *alb- (~ *albh-), *kar-/*kal-, and *borm- (> *bormo-) systems, have allowed us to assess the possibilities of language-contact phenomena in prehistoric ages among speakers from different language families, in the dynamics of Indo-European versus non-Indo-European and/or of Proto-Indo-European versus Pre-Indo-European.

We have also been able to focus on the important notions of ‘reuse’ and ‘re-functionalisation’ of roots and proto-forms in remote times, which are phenomena characteristic of linguistic contact. Due to the interaction among different speakers, a root or a lexeme in general acquires different meanings, morphology, and functions in the languages of the new speakers. This phenomenon can be prevalently reconstructed through the study and etymological analysis of ancient toponyms. Place names are, therefore, one of the few ‘windows’ we have to observe a linguistic world that is lost.

In the next chapter, we will turn our attention to contexts that do not have well-documented historical sources. This happens mainly in the case of undocumented languages and language families.