Part I
Overviews
1
Lineage and the constructive imagination
The birth of historical linguistics
Roger Lass
1 Precursors: the ideas of ancestry and change
My topic is too complex for the allotted word limit. So rather than aiming at exhaustiveness,1 I confine myself to two major intersecting themes: the notion of genealogy, and the possibility of reconstruction, with emphasis on the latter. I will therefore neglect at least two important topics: (a) the relation of linguistics to the life sciences; this includes the debate, especially prominent in the 1860s, over whether language is an âorganismâ and should be treated as a biological object rather than human action in society;2 and (b) the âromanticâ elements of the subject.3 This will rather be âgreat moments in early historical linguisticsâ, with the emphasis on those conceptual innovations that approach, or are ancestral to, what is now considered âmainstreamâ. This is not âWhig historyâ, evaluating past works according to how closely they approach the âgoalâ of being modern; rather an attempt at describing and contextualising the early days of our craft, and incidentally revaluing some early work and showing how much older some of our fundamental ideas are than we habitually think.
People have been conscious that language has a temporal dimension at least as long as they have been writing about it. In the West, perhaps the earliest âseriousâ recognition of language change is the collection of speculative etymologies and discussion of the meaning of letters, sounds and names in Platoâs Cratylus.4 One primary issue there is the essentialist question of whether words have meaning by nature or by convention. Semantic change and dialect difference are invoked as a partial argument for conventionality. There is also a claim of monogenesis by an act of creation â names were given by some figure in the distant past (the âLegislatorâ). Change and variation are then seen as betrayals of this original creation. The notion of âoriginalityâ reappears in different forms later on. This is a fairly isolated example; by and large language in time was not a focal concern in the Classical traditions, Greek or Latin, though language as a philosophical object certainly was.5
The focus however did change. In the later Judaeo-Christian tradition (itself profoundly historical)6 there was a topos which furnished the metaphoric and intellectual basis for a rich tradition of discussion: the story of Babel (Genesis 11). The primary concern was determining what language the people who built the Tower spoke before the dissolution into many mutually incomprehensible languages brought upon them as Jahwehâs punishment for their hubris. This was to remain central to scholarly discourse well into the Baroque, especially in Germany and the Low Countries (McLelland 2011). The essential question was: what was the original language of mankind? One typical early monogenetic identification by an authoritative figure is Isidore of Sevilleâs seventh-century assertion (Etymologiae 1.3.4) that Hebrew is to be seen as âomnium linguarum et litterarum ⌠matremâ (âthe mother of all languages and literaturesâ).7 I introduce this here to capture a very different way of thinking about the world, which slowly disappeared, but for a time continued alongside a more âmodernâ kind of thought. Monogenesis from Hebrew in particular was taken seriously by some as late as the seventeenth century, often in a cultural nationalist framework, i.e. attempting to derive oneâs own language from Hebrew and thus specially validate it.8 But the increasing growth of data collection (see below) gradually reduced these attempts to marginal speculation.
By the sixteenth century, monogenesis, though still discussed, became less important. The introduction of Semitic grammatical works into Europe seems to have stimulated the serious adoption of the idea of there being many families of related languages, which are nonetheless not related to other linguistic families.9 Thus polygenesis came to be a foundation concept. An early example is Theodor Biblianderâs work on Semitic (1548), which identifies Arabic as a descendant of Hebrew (this idea was already current in the ancient Semitic grammatical tradition), but not related to any European languages. The idea of families of independent origin, descending from unrelated âmotherâ languages, was commonplace:10 Gessner (1555) allowed for both mother languages that were unrelated and those that were related (cognatae).
By the seventeenth century we can see the outlines of what was to become the modern view of language filiation. Mother languages generate later dialects the way plants produce branches or shoots (Scaliger 1610: 119: âmulti dialecti tamquam propagines deductae suntâ). A more historical way of thinking (of a type now ordinary but then conceptually new) can be seen in the same work: Scaliger visualises an âoriginalâ persisting through time and space in changed forms, so that Italian genero, Spanish yerno, French gendre âson-in-lawâ < L gener could be said still to âbe Latin wordsâ, in a special historical sense.11 This is precisely the way we (some of us anyhow) might now identify the preterite of a Germanic strong verb (e.g. wrote, sang) as âan old o-grade perfectâ. Within English it is not one, not least because English does not have an ablauting perfect; but in the historianâs eye, in a very real sense, it is. Historians have a way of seeing things sub specie aeternitatis12
As more data from different languages became accessible during the âage of explorationâ (largely through grammars written by missionaries),13 the typological range of known languages increased hugely. The idea that there could be numerous unrelated âlinguae matricesâ, some of which were the roots of macrofamilies, became commonplace. Even what would later be called the Indo-European family, in something akin to its modern shape, had some currency as early as the sixteenth century. It lacked a reliable etymological foundation, being based mainly on phenotypic comparison bolstered by external history and even cultural likenesses, but there were convincing examples. Many scholars proposed a great âScythianâ nation once covering much of Europe and Asia. It thus included most of what we now term Indo-European, though some writers included Uralic as well. There was also a âCelticâ school, which posited a somewhat different great ancestral family. Possible macrofamilies, and the question of which languages were members of which ones, were seriously discussed through the later eighteenth century. This discourse finally blossomed (via the transformation of the ideas of Scythian and/or Celtic in the old sense) into what was increasingly perceived as something very like the modern idea of Indo-European.14
So it is important to note that the idea of such a language family was present in linguistic discourse long before Sir William Jonesâ famous address of 1786 (published 1788), where he made the statement that apparently must be quoted in any discussion of early historical linguistics or comparative method. This did not represent an explosive beginning to âmodernâ thinking, as is sometimes implied in textbooks; but it portrayed in a programmatic way the foundations of a new discourse, which was to dominate the field throughout the nineteenth century, and whose basic ideas still underlie all respectable historical linguistics.15 In his Third Anniversary Discourse âOn the Hindusâ, Jones said (1807 [1788], iii, 34):
The Sanscrit language, whatever be its antiquity, is of a wonderful structure; more perfect than the Greek, more copious than the Latin, and more exquisitely refined than either, yet bearing to both of them a stronger affinity, both in the roots of verbs and in the forms of grammar, than could possibly have been produced by accident; so strong indeed, that no philologer could examine them all three, without believing them to have sprung from some common source, which, perhaps, no longer exists [âŚ]
He also suggests that there are reasons for believing that âthe Gothick and âthe Celtick âhad the same originâ with Sanskrit, âthough blended with a very different idiom, and perhaps âthe old Persianâ as well.16 By the end of the century more detailed demonstrations of relationship, primarily on morphological grounds, had been established for Lapp (Sami) and Hungarian (Sajnovics 1770) and for Finnish, Estonian and others (Gyarmathi 1799). Because of the languages dealt with, and in part because of âE...