Historical linguistics investigates the historical development of languages and the relationships between different languages. By reconstructing the spread of languages, the influences between languages, and the meaning of words in languages, scholars can (amongst other things) argue for the existence and use of particular concepts in different cultural groups. Languages correspond closely to ethnic groups and we can deduce important facts about social practices, beliefs, and identity through historical linguistics.
The Celtic language family is a branch of the Indo-European language family. Linguists are still determining the relationship between the branches of Celtic languages; while some believe that there is an early and fundamental division between Insular (meaning, “belonging to the British Isles”) and Continental languages, others believe that the division lies between P-Celtic and Q-Celtic branches.
Place names, as recorded in early documents, help us determine the language that was spoken by the people who coined those names. Place names can thus act like fossils recording the existence of people long after they and their language disappear, and they sometimes also offer some insight into beliefs and practices.
Literacy refers to the ability of people to read and write. Like many other people, various Celtic peoples learnt literacy from their contact with the mediterranean peoples. Different Celtic regions began to use different scripts depending on which literate people they were in contact with. Most surviving texts written by Celtic peoples themselves tend to be rather short inscriptions written on burial stones or ritual sites. Occasional words and names survive in the Greek and Latin authors as well. It is only in the early medieval period that we start getting extended texts written by Celtic peoples themselves in their own languages.
Languages are dynamic entities, almost like biological organisms: the sounds, words, and rules of the language spoken by communities change over time. Sometimes this is due to influence from other languages and social groups, but it is also natural for the speakers of a language, even when they are isolated, to create new words, forgot old ones, and change the way that they pronounce words and form sentences. It is as though the “DNA” of a language can evolve over time.
Languages are also similar to biological species in that there are “genetic” relationships between different languages, which can be placed in families of languages which are derived from a common ancestor.
The Indo-European Language Family
Since the early nineteenth century, scholars have been systematically analyzing a set of languages which have been classified as belonging to the “Indo-European” language family, because (as of the medieval period) they covered a broad range of territory of Europe and Asia, as far west as Ireland and as far east as India. We could place the major families within the Indo-European language family on a map (according to where they were at the beginning of the Common Era) thusly:
There have been many attempts to create detailed family trees of the various branches of the Indo-European language families. Although a tree structure (like a family tree for human families or the evolutionary tree of biological species) oversimplifies the complex relationships between these different language branches, it is still a useful way to conceptualize the differentiation of “child” languages from a common “parent.” The first real pioneer of Indo-European linguistics, August Schleicher (1821-68), proposed a family tree to represent the relationships between the “children” of the Proto-Indo-European “parent” language; although this language model is now obsolete, it is a good starting point for learning about I-E linguistics.
English, Dutch, and Swedish are descendants of the Germanic family of languages; the Italic family includes Latin, French, Spanish, Italian, etc.
A look at a selected list of words which represent numbers from Indo-European languages demonstrates that these come from a common source, such as the words for these numbers:
Contrast the Indo-European languages with the words for numbers in Hebrew: s(e)nayim “two,” `arba`a “four,” sib`a “seven,” `asara “ten,” me׳a “hundred.” There is no systematic resemblance between the Indo-European words and the corresponding Hebrew words because Hebrew is a member of the Hamito-Semitic language family, not the Indo-European language family.
If Indo-European languages come from a common source, what accounts for changes to words between languages? How do languages evolve? To return to our biological metaphor: What kind of “mutations” are there and what causes these “mutations”? Some of the most common kinds of mutations are:
- The meaning of words change
- The pronunciation of certain sounds change
- The sounds within a word can change places
- Some sounds are added or left out of the pronunciation of words
These principles can be illustrated with examples from English. There are many ways in which a word can change meaning. It can become more narrow in meaning: while in North America “school” refers to any educational institution, it is used exclusively for primary and second levels in Britain (not of colleges or universities); the Old English word dēor referred to any wild animal, but become restricted later in meaning to “deer.” Words can become more broad in meaning: the Middle English estoffe was used specifically of material used for making clothing, but has evolved the general meaning “stuff.” Words can also acquire secondary associations and then shift towards those associations: the word “music” is ultimately derived from the Greek phrase mousikē (tekhnē) “the (art) of the Muses,” as the Muses were believed to be the source of creative inspiration.
Some sounds – individual vowels, or consonants, or clusters – can change. During the development of Old English from its Western Germanic origins the consonant cluster “sc” in Old English became “sh”: thus Germanic scip became “ship,” Germanic fisc became “fish,” and so on. Between the fifteenth and the eighteenth century the entire system of vowels in English (at least as spoken in southern England) underwent a systematic set changes, usually referred to as the “Great Vowel Shift.” The word “house,” for example, used to rhyme with the word “moose,” as we now pronounce it (and in fact, many of the old vowels are preserved in conservative dialects of “Scots,” which is closely related to Middle English).
Sometimes two nearby sounds change places in a word (a phenomenon technically referred to as metatheses). This is especially common for the consonants called liquids (the letters “r,” “l,” and “n”). For example, in some dialects of English “pretty” is actually pronounced as though it were spelled “purdee”; the word “comfortable” is actually pronounced as though it were spelled “cumfterbul”; and George Bush will long be remembered for pronouncing “nuclear” as though it were spelled “nuculer.”
Sometimes sounds are left out of the pronunciation of a word. In common speech, the word “every” is usually pronounced as though it were spelled “evry.” The “gh” in words like “light” and “night” used to be pronounced in early Middle English, and are still pronounced in conservative dialects of Scots. However, this sound was later omitted but in compensation, the vowel was “lengthened” (technically speaking, the vowel became a diphthong). Similarly, the final “-e” in most words stopped being pronounced in words like “name” and “line” during the Middle English period, with a compensatory lengthening of the internal vowel.
Linguists have reconstructed many words in the ancient Proto-Indo-European parent language, even though there are no written records of this language, simply by comparing surviving words in the languages descended from it, inferring the system of sound changes that happened over time, and working backwards. To indicate that a word has been reconstructed, but does not survive in any written record, it is preceded by an asterisk “*”. The less-than “<“ and greater-than “>” symbols are used in shorthand notation to indicate that one word is derived from another.
Over long periods of time, combinations of changes can completely alter a word so that it is hardly recognizable. To return to an ancient Indo-European example, the root PIE *(s)teg- “cover” developed separate forms and usages in derived languages: Latin toga (< tegō “I cover”), Old High German dah “roof,” Lithuanian stógas “roof,” Old Gaelic tech “house.” Words derived from a common “ancestral” word, as all of these words are from *(s)teg-, are called cognates.
Studying the vocabulary of a language – its lexicon – is useful for a number of reasons. First of all, the existence of a word in language implies that a corresponding concept or item existed for the speakers of that language. For example, the fact that all of the branches of the Indo-European family have a word for “dog” that derives from the same root tells us that the dog had been domesticated and known to the people who spoke the ancestral Indo-European language. This is suggested by examples of related words in several Indo-European languages (note how the earlier Indo-European “c” sound in this word became “h” in the Germanic languages):
The fact that the words for “wheel,” “axle,” and “hub” can be reconstructed from all of the Indo-European languages implies that the common ancestral language, and therefore culture, had wheeled transport, either a chart or wagon. On a more abstract level, all of the branches of Indo-European have words that mean “sister’s son” and “mother’s brother,” implying some ancient and important social structure and function between these two members of a family.
Studying the lexicon of a language can also provide valuable clues about contact between peoples in prehistoric times which would be impossible to discover in any other way. Let’s say that we find the word “A” has been borrowed from language “X” into language “Y”: this implies that the societies that speak X and Y were in contact at some point. There are a number of different reasons why Y may have borrowed A: perhaps the concept or item corresponding to A did not exist previously in Y (at least in some specific form); Y may have found the word A useful to adopt in addition to (or instead of) some similar word they already had; or perhaps X put Y under some kind of duress to adopt A.
Let’s look at a few historical examples from the English language. When the Anglo-Saxons adopted Christianity, they borrowed many Latin terms to describe the concepts of their newly adopted religion and its equipment: Latin presbyter developed into “priest”; Latin monachus developed into “monk”; Greek apostolos “messenger” developed into “apostle.” On the other hand, sometimes native Germanic words were shifted in meaning to suit Christian concepts, as when the Germanic gast “demon, evil spirit” was altered to refer to “soul” or the “Holy Ghost.”
After the Battle of Hastings in 1066, French-speaking Normans took over the top echelons of English society. This introduced a whole new influx of French words into the English vocabulary, especially at the higher ranks of society, who continued to speak French in “polite” company for about a hundred years. One result of this linguistic “layering” of society, with French speakers at the top and English speakers below them, were pairings of words with similar meanings, one a lower-status word of Anglo-Saxon origin and the other a higher-status word of French origin. The classic examples relate to food: pig / pork, chicken / poultry, calf / veal, and cow / beef. There are also legal and political terms, such as goods / chattels, lands / tenements, and fit / proper.
The Celtic Language Family
This map illustrates the locations of the main branches of the Celtic family around the beginning of the Common Era:
The history of these individual branches can be described as follows:
- A number of inscriptions have been found in the southern Alps, associated with an archaeological culture group known as the Golasseca culture. The language has been called Lepontic and has been accepted as a branch of the Celtic language family, although there is still debate as to whether it is an independent branch of the Continental Celtic languages or a dialect of Gaulish. The inscriptions date to as early as the late seventh century BCE. Lepontic may have been absorbed into the Gaulish dialects brought by Gaulish immigrants around the Po valley. All Celtic languages in Cisalpine Gaul became extinct after the Roman conquest of the region, between the second and first century BCE.
- The first textual evidence of Gaulish is from the sixth century BCE, mostly placenames and personal names from Greek sources. Our earliest inscriptions in Gaulish use the Greek alphabet and date from the third century BCE. Increasing Roman influence brought the use of the Latin alphabet in the first century BCE. Gaulish may have survived as late as the sixth century CE in isolated areas of Gaul.
- Galatian was the name given to the Celtic language of the Gaulish nations who went to Galatia in Asia Minor in the third century BCE. It seems to have survived until at least the fourth century CE.
- The earliest evidence of Celtiberian dates from only the late third or second century BCE. The conservative nature of the language suggests that it branched off from the Proto-Celtic ancestor at an early stage and is more akin to Goidelic than the Gallo-Brythonic branch. The earliest inscriptions use the Iberian writing system, but the Latin alphabet was used after the first century BCE. The Celtiberian language seems to have become extinct in the second century CE with the dominance of Roman influence.
- Some scholars use the term Hispano-Celtic for other branches of Celtic found in the Iberian peninsula, especially in the west and north, for which there is now only fragmentary and inconclusive evidence.
- The Insular languages are those languages spoken in the British Isles (Britain and Ireland). Insular means (in this context) “of the islands.”
- Goidelic is the branch of the Celtic languages spoken in prehistoric Ireland and probably in Argyll (Scotland). The earliest inscriptions in Goidelic date were written in the Ogam script and may date to as early as the third century CE. Goidelic developed into distinctive branches in Ireland, Scotland, and the Isle of Man, which are still living languages in all of these places.
- Brittonic (sometimes also called “Brythonic”) is the branch of the Celtic languages spoken on the island of Britain. Especially due to the influence of the Roman Empire, and the divisions imposed by a series of conquests and occupations, Brittonic developed into distinct dialects, including Pictish, Old Welsh, Cumbric, and Cornish (Breton is derived from Cornish).
There is still considerable debate about the relationships between the various members of the Celtic language family. This is largely due to the fact our evidence for some members of the family – especially ones killed off by the Roman Empire – is very sparse and scanty. The two most popular models at present are represented by these family trees:
The Goidelic and (Gallo-)Brittonic branches are often alternatively referred to as Q-Celtic and P-Celtic (respectively) because of one of the characteristic sound changes that evolved in the prehistoric period: the “kw” (or “qu”) sound in Indo-European was retained in early Goidelic but became “p” in later branches of Celtic. This change is not unique: a parallel sound change happened in branches of Greek. The “q” sound as seen in Latin equus “horse” changed to a “p” sound in Greek híppos.
The fact that Celtiberian is a Q-Celtic language (like Goidelic) demonstrates that it split off of Proto-Celtic before the development of kw > p in the (Gallo-)Brittonic branches. In other words, Celtic speakers existed in Ireland and Iberia (whether because they migrated there or Celtic evolved there, a question to be raised later) before the P-Celtic branch evolved (this has important historical implications explored further in later units).
We can see the kw > p shift in Goidelic and Brittonic word pairs such as:
|crann “tree, branch”
There are several key sound shifts that characterize the sound changes in the Celtic language family: the “p” sound in Proto-Indo-European was lost altogether (indicated by the formula PIE *p > PC ∅) and the Graeco-Italo-Celtic vowel ē changed to ī. Here are some examples comparing Celtic forms with the reconstructed Proto-Indo-European, Modern English, and Latin forms (note that some of the meanings of words shifted significantly):
|*porḱo- “litter of pigs”
|farrow “litter of pigs”
|hrif (e.g., “midriff”)
What seems to have happened to PIE *p in the Germanic languages?
As mentioned previously above, place names can act like fossils, providing testimony for the existence of speakers of languages in particular areas long after they disappear (or stop speaking their language). The ghostly echoes of First Nations linger over the landscape of North America long after they or their languages have been been lost to European settlement: think of “Ohio” (an Iroquoian name), “Quebec” (an Algonquin name), or “Antigonish” (a Mi’kmaw name).
Similarly, Celtic names have persisted in many parts of Europe despite the fact that the Roman Empire assimilated the Celts in the early centuries of the Common Era. This wide distribution of Celtic names attests to their presence in prehistoric times. Of the various features of the landscape, the names of rivers tend to change the slowest, and hence the names of rivers usually give us the oldest possible linguistic layer. For example, the river Rhine and all of its eastern tributaries – the Neckar, Main, Luhn, Ruhr, and Lippe – are in origin Celtic names.
Here are some of the most common Proto-Celtic elements in European place names:
|“hill, high place”
|Eburobriga “Yew-hill” > Avrolles, France
|“crooked, bent; river bend”
|Cambodunum “fort of the river bend” > Kempten, Germany
|“enclosed site, fort”
|Lugudunum “Lug’s fort” > Lyon, France
|“large settlement, oppidum”
|Durocornovium “City of the Cornovii people” > Wanborough, England
|“plant”; OG “yew”, MidW “cow parsnip”
|Eburobriga “Yew-hill” > Avrolles, France
|“plain”; “settlement on plain”
|Argantomagus “Silver city (on the plain)” > Argentan, France
|Drunemeton “Oak’s sacred place” in Galatia, Turkey
|Segedunum “Strong fort” > Wallsend, England
It is sometimes possible to find essentially the same place name in slightly different forms in multiple Celtic countries: for example, Vindo-magus [Gaulish “White-plain”] in Gallia Narbonensis corresponds directly to Gwynfa [Modern Welsh] in Montgomeryshire, Wales, and Findma [Old Gaelic] in County Antrim, Ireland.
Personal Names and Ethnonyms
Names are important: they convey values, beliefs, and ambitions, and thus open a small window into the lives of people. Modern Gaelic and Welsh words can often help us to interpret the names of many ancient Gaulish and British chieftains and nations:
|Welsh and Irish Elements
|domain “world” + ríg “king”
|aingid “protects” + mór “great”
|ecraide “horse herd” + ríg “king”
|bet “world” + ríg “king”
|for “over, above” + cingid “steps, walks” + ríg “king”
|aile “other” + bro “territory”
|brig “high; exalted”
Translate the following Celtic place names, making use of the information about Celtic word meanings above:
- Segobriga (Segorbe, Spain)
- Vindobriga (Vienne, France)
- Rigodunon (Castleshaw, England)
- Nemetoduron (Nanterre, France)
Clackson, Indo-European Linguistics.
Crystal, The Stories of English.
Dillon and Chadwick, The Celtic Realms.
Koch, An Atlas.
—, “Interpreting Tartessian.”
Mallory, In Search of the Indo-Europeans.
Matasović, Etymological Dictionary.
Russell, An Introduction.