More on geographic determinism and language
In a recent post, I’ve discussed the issue of geographic determinism, that is the hypothesis that certain kinds of terrain or weather favor certain structural features in languages. As I have shown there, this simple-minded version of geographic determinism – e.g. thinking that cold/damp climatic conditions lead to the appearance of nasal vowels – does not work. However, it is not entirely true that linguistic structures are completely independent of the location where the given language is spoken. In other words, a given language transposed to another place on Earth (or to another planet) would or could be exactly the same. Quite the contrary: certain structural features in language depend on location and terrain.
In her pioneering work on this subject, Johanna Nichols of UC Berkeley has distinguished spread zones and residual zones, subcontinental areas whose definitions are based on relative diversity, center, periphery, and internal stability (cf. her 1992 book Linguistic Diversity in Space and Time). Spread zones are characterized by low levels of linguistic diversity (i.e. few language families and much similarity even across family lines), relatively young age of language families and a rapid spread of languages (and families) across the zones. The concept of spread zones is illustrated best by the Eurasian steppes, the area that serves as home to only two large language families: Turkic and Mongolic languages. Other prototypical spread zones considered by Nichols in her book include western Europe, central Australia and interior North America; according to Nichols,
“The first two are prototypical spread zones dominated by a single family (Indo-European and Pama-Nyungan respectively). Interior North America is mostly covered by a few language stocks [her term for language families – Asya], notably Algonquian and Siouan.”
The list of spread zones may also include Mesoamerica, the Ancient Near East, central insular Oceania (home to Austronesian languages), central sub-Saharan Africa (where the Bantu family has spread widely), the Great Basin of North America (where the Numic branch of Uto-Aztecan has spread widely and rapidly) and the entire Arctic region (home to the Eskimo language family).
In contrast, residual zones are distinguished by high levels of linguistic diversity (i.e. numerous language families and high levels of structural diversity), older age of language families and no appreciable spread of languages and families over territory. In simple terms, residual zones are far more mosaic than spread zones. The Caucasus region provides an excellent example of a residual zone: it is home to three (potentially four, depending on analysis) families, as well as languages from other language families, such as Ossetian and Armenian (both Indo-European), Balkar and Azeri (both Turkic). Nichols’ list of residual zones includes: a part of eastern Africa including Ethiopia and Kenya; the Pacific coast of northern Asia (from Japan to the Bering Strait), northern Australia, the Pacific coast of North America (especially California), parts of South America, New Guinea and the Pamir-Himalaya region.
A quick comparison of the two types of geolinguistic zones reveals an interesting correlation: spread zones are typically flat, easily-accessible areas, whereas residual zones correlate with areas of complex topography, often mountainous, but in other cases with rugged coastlines, clusters of smaller islands, marshy areas, tropic jungle and so on.
But the distinction between spread zones and residual zones correlates not only with geography but also with structural features of languages. For example, Nichols shows that the property of head-marking is associated more closely with residual zones, while dependent-marking is more typical in spread zones. Taken to apply on the clausal (sentential) level, the distinction between head-marking and dependent-marking languages amounts to whether overt morphosyntactic marking reflecting the syntactic relations (e.g. subject, object, etc.) is located on the verb (i.e. head of the clause) or on noun phrases (i.e. on dependents). (There are also languages that do both or neither, but I will disregard them here.)
An example of a head-marking language is Abkhaz (a Northwest Caucasian language of some 105,000 speakers), where prefixes on the verb are used to encode agreement with the subject, the object and the indirect object. In a-xàc’a a-pħwəs a-šq’wə ø-lə-y-te-yt’ ‘The man gave the woman the book’ (literally ‘the-man the-woman the-book it-her-he-give-PAST’), the prefix lə- shows agreement with the indirect object ‘the woman’, the prefix y- — with the subject ‘the man’, and the absence of another prefix (marked by ø-) — with the direct object ‘the book’.
Dependent-marking languages can be illustrated with Uradhi, a Pama-Nyungan language, a close relative of Dyirbal, once spoken in Queensland, Australia. In this language, subjects and objects bear cases marking; in the example below, the subject ‘old man’ is marked by the ergative suffix -nku and the object ‘firewood’ – is marked as absolutive (which has a zero ending, but as this is the only zero ending in the case paradigm it is unambiguously identifiable as absolutive, hence it counts as a case). The verb has no agreement with either subject or object; the only bound morpheme attached to the verb is the past tense suffix -n.
|‘The old man picked up some firewood.’|
As the map below shows, head-marking languages are most common in such residual zones as New Guinea, parts of South America and the Pacific coast of North America, whereas dependent-marking is more common in such prototypical spread zones as western Europe and central Australia. Thus, in effect geography does determine, albeit indirectly, some structural features of a language, so geographical determinism cannot be discounted completely.
Like this post? Please pass it on:
« On geographic determinism and nasal vowels...
Iced tea or ice tea? »