Editing Ligurian language (ancient) (section)

== Classification ==
Most experts on the ancient languages of the region, including [[Patrizia de Bernardo Stempel]], [[Xavier Delamarre]], [[Jürgen Untermann]], [[Javier de Hoz]], Franceso Rubat Borel, and Bernard Mees, agree that Ligurian was part of the [[Indo-European language|Indo-European]] language family.<ref name="consensus" />

However, there is debate regarding the relationship with [[Celtic languages|Celtic]], as Ligurian exhibits both similarities and differences with neighbouring Celtic languages (i.e. [[Gaulish]] and [[Lepontic language|Lepontic]]).{{Sfn|Mees|2024|p=|pp=203–204, 209}} De Bernardo Stempel has proposed that Ligurian might represent an archaic Celtic dialect that was influenced by a non-Celtic [[Substratum (linguistics)|substratum]] to account for these differences,{{sfn|de Bernardo Stempel|Arenas Esteban|2011}}{{Sfn|Mees|2024|p=|pp=203–204, 209}} and Delamarre believes "that the term 'Ligurian' refers to the population involved in the first Celtic invasions into southern Europe".{{Sfn|Delamarre|2007|pp=36–37}} Conversely, Rubat Borel, de Hoz, Untermann and Mees argue that Ligurian was a separate Indo-European language or language family, either related to or influenced by Celtic.{{Sfn|Mees|2024|p=|pp=203–204, 209}}{{Sfn|de Hoz|2005|p=175}}

There is also debate about the geographical range of linguistic elements that can be attributed to the Ligurian language.{{sfn|Mees|2003|pp=16–18}}{{sfn|Clackson|2015|pp=3–5}} Javier de Hoz has proposed distinguishing between the region inhabited by the 'Ligures' of the Roman ''regio Liguria'' (between [[Ventimiglia]] and [[Ameglia]]), and the area mentioned in early ancient sources as home to the 'Ligues' (from the [[Arno|Arno River]] to west of the [[Rhône]], and possibly even as far as northeastern Spain). He refers to the language of the former as 'restricted Ligurian', noting that this group had a well-defined cultural identity and likely used an Indo-European language about which little is known. In contrast, the latter region, which he described as 'broad Ligurian', was inhabited by culturally related but likely diverse peoples who may have spoken various languages.{{Sfn|de Hoz|2005|p=175}}

=== Relationship with Celtic ===
Linguists [[Patrizia de Bernardo Stempel]] and [[Xavier Delamarre]] have argued that many names of tribes described by ancient scholars as 'Ligurian' can be explained as Celtic. This includes the ''[[Ingauni]]'' (<small>Celt.</small> *''Pingāmnī''), ''[[Salyes]]'' (< *''Sḷwes''), ''[[Ligauni]]'' (<small>Celt.</small> ''līg''- or ''liga''-), ''[[Intimilii]]'' (<small>Celt.</small> ''uindi-''), [[Marici (tribe)|''Maricii'']] (<small>Celt.</small> ''maro-''), and ''[[Oxybii]]'' (<small>Celt.</small> ''oxso''- or ''uxso''-),<ref>{{harvnb|de Bernardo Stempel|2006}}; {{harvnb|de Bernardo Stempel|Arenas Esteban|2011}}</ref> as well as the ''[[Iemerii]]'' (<small>Celt.</small> ''iemur-''), ''[[Orobii]]'' (<small>Celt.</small> ''orbi-''), ''[[Segobrigii]]'' (<small>Celt.</small> ''sego-'') and ''[[Reii]]'' (<small>Celt.</small> ''*riio-'').{{sfn|Delamarre|2003}}

According to de Bernardo Stempel, such linguistically Celtic tribal names suggest that a 'Celto-Ligurian' dialect played an important role among the languages spoken in ancient Ligury.{{Sfn|de Bernardo Stempel|2006|p=46}} Furthermore, she notes that some lexical items appear to be common to Ligurian and Celtic, such as ''cotto''- (''[[Alpes Cottiae]]''), ''gando''- (''Gandovera''), ''ambi-'' (''pago Ambitrebio''), ''ebu''- (''Eburelia''), ''medu''- (''Medutio''), ''seg''- (''Segesta Tigulliorum''), ''catu''- (''Catucianum''), and ''roud''- (''Roudelium'').{{sfn|de Bernardo Stempel|Arenas Esteban|2011|pp=129–130}}

Arguing for a connection between Ligurian and Celtic languages, de Bernardo Stempel has listed the following isoglosses as common traits shared by both language groups:{{sfn|de Bernardo Stempel|Arenas Esteban|2011|pp=129–130}}

* IE *''bʰ'' > ''b'', as shown by ''Comberanea rivus'' and ''Badiennon'', from the Celtic ''badius'' 'shiny; yellow, blond'
* loss of initial ''p''-, as in ''Ingauni'' < *''ping-amn-ī''
* *''l̥'' > ''al'' before resonant, as in ''*Sl̥wes'' > ''Salues''
* lenition of voiced consonants, as in ''Ligauni'' < *''ligamni''
* unstressed *-''i̯o''- > -''je''-, as shown by ''Nitielium'' and ''Berigiema''
* ''o'' > ''u'' before labials, as in ''Leucumellus'' and ''Latumarui''
* assibilation, as shown by ''Mezu'' and ''Meśiolano''
* palatal anticipation, as in ''Airuno'', ''Airasca'', ''Airolo'', and ''Eluveitie''
* epenthesis, as shown by ''Berigiema'' and ''Alebinna''
* suffixes *-''enko''- and *-''asko''-, as in ''Bodincos'' and ''Vinelasca'' (< *''wind-el-askā'')

Amongst Celtic subgroups, common innovations are only shared with [[Lepontic language|Lepontic]], including *''ks'' > ''s'', *''nd'' > ''n(n)'', *''st'' > ''z'', ''VCi̯V'' > ''VjCV'', the analogical acc. pl. *-''aś'' > -''eś'' after the nom. pl. -''es'' (consonantal stems), gen. sg. -''j''-''os'' & -''ei''-''s'' → -''ei'' -''os'' (i-stems), gen. sg. *-''osjo'', 3rd sg. preterite in -''te'', patronymic -''alo''-, patronymic -''ikno''-, gamonymic -''iknā'', and monothematic personal names. Conversely, some innovations are shared between Ligurian and the [[Gaulish language]] of the early sources, such as gen. sg. *-''ī'', 3rd sg. preterite in -''tu'', and patronymic -''ikno''-.{{sfn|de Bernardo Stempel|Arenas Esteban|2011|pp=129–130}}

=== Non-Celtic Indo-European forms ===
On the other hand, some Ligurian ethnonyms show a weaker or less evident connection to Celtic, such as the name ''[[Friniates]]'', which can be hardly regarded as genuinely Celtic.{{Sfn|Mees|2024|p=207}} Although ''[[Deciates]]'' seems to derive from the [[Proto-Indo-European]] (PIE) root *''deḱ-'', it does not appear to be Celtic. Consequently, [[Javier de Hoz]] suggests classifying the name as 'restricted Ligurian', given the tribe's geographical location.{{sfn|de Hoz|2005|pp=178, 181}} The ethnic name ''[[Taurini]]'', from the Indo-European *''tauros'' 'bull', does not follow the [[Metathesis (linguistics)|metathesized]] Celtic form ''taruos''. According to Delamarre, this could be explained by the influence of Latin or the preservation of an archaic form.{{sfn|Delamarre|2003|p=291}}

Similarly, the Ligurian ''Bodincos'' 'of unmeasured depth', from PIE ''*bʰudʰnós'', does not display the metathesis of [[Italo-Celtic]] *''bʰundʰós'' (cf. Latin ''fundus'', Middle Irish ''bond'').'''{{Sfn|Mees|2024|p=208}}''' The ethnonym ''[[Eguiturii]]'' (< PIE *''h₁éḱwos'' 'horse') also features an archaic preservation of [[Labialized velar consonant|labio-velar]] ''-kʷ-,'' in contrast to Gaulish ''epos'' (which can be compared to the Ligurian ''[[Epanterii]]'').{{Sfn|Evans|1967|p=200}} The similarity between the ancient names of Genoa (''Genua''), in the ancient Ligurian region, and Geneva (''Genaua''), in Celtic-speaking territory, has been highlighted by scholars,{{Sfn|Delamarre|2007|pp=36–37}} but if both toponyms could derive from the Celtic ''*genu'' 'mouth', the PIE stem *''ǵónu''- 'knee', which is commonly found in other Indo-European language groups, may also be at the origin of the names.{{sfn|Clackson|2015|pp=3–5}}

Francesco Rubat Borel has highlighted unusual Ligurian phonological changes not seen in Celtic languages, such as *''upo''- > ''uea''- and *''o'' > ''a'' in unstressed syllables, the toponymic suffix -''asc''-, and the retained *''p'' in ''Porcobera'', which would make claims of Ligurian as a Celtic language problematic.{{sfn|Rubat Borel|2008}}{{Sfn|Mees|2024|p=209}} According to Bernard Mees, Ligurian appears to share some phonological features with Celtic while differing significantly in morphology, suggesting it was a separate language with certain similarities to Celtic.{{Sfn|Mees|2024|p=|pp=203–204, 209}} He writes that "the Romans did not consider the Ligurians to be Celts, and although there are some evident parallels between Ligurian and Celtic names, these can all be understood as reflecting inherited Indo-European features, not distinctively Celtic traits."{{Sfn|Mees|2024|p=220}} By analysing Ligurian personal names, [[Jürgen Untermann]] concluded that Ligurian is best considered an Indo-European language or language family separate from both Celtic and Italic.{{Sfn|Untermann|2006|pp=1762–1766}}{{Sfn|Mees|2024|pp=203–204, 209}}

=== Substrate theories ===
In the late 19th century, [[Henri d'Arbois de Jubainville]] proposed that the Ligures constituted an early Indo-European substratum in Western Europe, separate from both Gaulish and Italic groups. Building on classical sources, he identified a range of place names and tribal names that did not fit neatly into known Celtic or Italic patterns. His theory, which came to be termed "Celto-Ligurian", influenced philological and archaeological approaches for the following decades. According to Bernard Mees, de Jubainville became "the intellectual grandfather to a genealogy of prehistorical and protohistorical substratum theories".{{sfn|Mees|2003|pp=16–18}}[[File:Old_European_hydronymic_map_for_the_root_*al-,_*alm-_Krahe.jpg|right|thumb|400x400px| [[Hans Krahe]]'s "[[Old European hydronymy|Old European]]" hydronymic map for the root ''*al-'', ''*alm-.'']]By the 1920s, scholars were using the "Celto-Ligurian" idea to explain problematic Indo-European toponyms and hydronyms across much of Europe. For instance, [[Paul Kretschmer]] argued that some inscriptions in Etruscan script (later identified as [[Lepontic language|Lepontic Celtic]]) provided evidence for a Ligurian linguistic layer, but subsequent discoveries established these as clearly Celtic with only limited Etruscan influence. In the 1930s, [[Julius Pokorny]] adopted these insights for his [[Pan-Illyrian hypotheses|pan-Illyrian]] (or "Illyro-Venetic") theory, linking it to the prehistoric [[Urnfield culture]]. Consequently, many difficult place-name etymologies were attributed to a hypothetical Illyrian layer, leading to broad, stratigraphical theories that traced Indo-European linguistic influences from Gaul all the way to the Balkans.{{sfn|Mees|2003|pp=16–18}}

By the late 1950s, Pokorny's theories had lost its momentum following critical scrutiny. The underlying place-name elements championed by de Jubainville and Pokorny, however, were reworked by [[Hans Krahe]] into his "[[Old European hydronymy|Old European]]" theory. Focusing on hydronyms, Krahe advanced a more refined approach, yet it remained conceptually indebted to de Jubainville's earlier "Celto-Ligurian" framework. Though Krahe proposed a more systematic argument than the earlier "Illyrian" or "Celto-Ligurian" frameworks, his theory still faced criticism for assuming that widespread, older Indo-European features belonged to one single language rather than several archaic dialects.{{sfn|Mees|2003|pp=16–18}}

Linguist [[James Clackson]] has criticized these approaches by stating that "the label 'Ligurian' merely serves to conceal our ignorance" about the pre-Roman linguistic landscape in various regions of Europe.{{sfn|Clackson|2015|pp=3–5}}