In many parts of the world, extended contact among speakers of different languages has led to the development of similar structures independent of genetic relations. The languages may be related to one another, but as a group those in a linguistic area (or Sprachbund "language union") show special properties not attributable to common inheritance. This can there be an issue in determining whether languages are truly related.
We'll look at a few examples of Balkan linguistic characteristics (Balkanisms), especially in syntax, before examining similar phenomena in the Americas. See a more complex map of ethnic groups (often defined by language) in the Balkans, found at this site. |
Many languages have a morphological future expressed using a suffix, in a way similar to the past tense, as in these Latin examples.
In the Balkans, however, we find a periphrastic pattern, where a separate word or other element combines with a verb to yield this meaning. Just as in English will, the usage derives from the meaning "want".
The two Greek forms are "I will see / be seeing". |
The definite articles in the Balkans are usually postposed, i.e. placed after the noun, even though the elements may derive from words that were originally placed before the noun (as in the Romanian article).
This is not true of every Balkan language; for example, Greek does not follow this pattern. |
For Uto-Aztecan especially, which extends all the way to Oregon, the northern languages do not share these traits. There are also smaller families or isolates not given in this list, such as Tequistlatec cited below. This linguistic area corresponds to a more general culture area with much shared history, which of course is the means by which linguistic traits undergo diffusion. |
The use of body-part expressions that we've mentioned before, such as "belly" for "in", is widespread in this area.
These words are not related — nor are the languages — but the metaphorical basis and the syntactic context have diffused across language boundaries. |
Mixtec, Kaqchikel: Campbell et al. (1986: 549); Nahuatl: Stolz & Stolz (2001: 1544)
While decimal numerals as in English are common (due to the fact that we have 10 fingers), other languages use base-5 (i.e. the fingers on one hand) or base-20 (i.e. fingers and toes together). Base 20, or vigesimal, systems are found throughout the Mesoamerican area. The word for 20 often involves the word "man" (i.e. a person's worth of digits). These examples are from Tequistlatec of Oaxaca, Mexico, a small isolate group (sometimes classified as Hokan).
The vigesimal pattern is ancient; numerals in the Mayan writing system (shown at right) were constructed from symbols that stand for 1 or 5, without a special symbol for 10; the cycle went from 0–19 and then restarted. The Celtic languages also had vigesimal counting, and this is the source of the standard French numerals soixante-dix (sixty-ten) for 70, quatre-vingt (four-twenty) for 80, and quatre-vingt-dix for 90. Some dialects have decimal numerals here, for example Swiss French septante, huitante/octante, nonante (based on 7, 8, 9). |
Campbell et al. (1986: 546)
The main culture areas in North America are shown in this map; these areas, or subparts of them, often constitute linguistic areas as well. |
The Northwest Coast is famous for its large inventories of complex sounds. For example, here are the consonants of Nootka. See also a map with languages named.
At the same time, certain sounds (such as /r/, and sometimes most labials) are absent from these languages. These patterns are reflected in the several different linguistic families represented in the region — Salish, Wakashan, Na-Dené, Penutian (or its subparts). |
The Bella Coola language of British Columbia is a good example of a language that permits long strings of obstruents. It's not clear how these strings should be analyzed in terms of syllables; many analyses have been proposed.
|
Alsea, located toward the southern end of this culture area, also permits complex strings of consonants (obstruents and sonorants), although not so impressive as Bella Coola.
Morpheme boundaries are included here. The longer words in Bella Coola are also morphologically complex; in fact, it has been claimed that a root morpheme in that language cannot contain more than 4 obstruents in a row (among other restrictions), as in /p̓xʷɬt/. "bunchberry". |
For example, Kashaya has this contrast in pairs of words such as these:
Langdon and Silver (1984) suggest that the distinction may have arisen from contact between languages with different realizations of the single /t/ in the respective languages; that is, some were dental and some were alveolar, and a contrast between the two may not originate in a single language but rather from the combination of languages. More specifically, the most common realization was probably alveolar but contact with dental languages could have led to the introduction of that segment. For example, some languages of the south have an allophonic alternation between the two places of articulation, which might have been reinterpreted by speakers of other languages. Alternatively, Catherine Callaghan reconstructs the contrast for the Miwok-Costonoan family and suggests it may have spread from there to other languages. |
Sounds that are similar to /s/ may have different realizations, and multiple s-like sounds (sibilants) may be contrastive in some languages.
Borrowings of some words with Old French s into Middle English with sh suggest the difference in pronunciation indicated here (other examples are the many words from French -ir verbs, stems in -iss-, that have English -ish: finish, polish, establish, etc.). However, many examples of Old French s were nevertheless borrowed into English as /s/, and certainly we read c(e,i) with /s/ in borrowings such as cent. Several possible reasons: the actual Old French sound was intermediate between English /s/ and /š/; the orthography played a role; or not all dialects of Old French had the same (or any) distinction. |
An issue in California reminiscent of the /t, ṭ/ issue is the plain and retracted sibilants. As Bright (1978) discusses, many English-speaking field-workers in California and some nearby areas have had trouble categorizing the sibilant sounds that they heard in native languages, often vacillating between transcriptions such as [s, ṣ, š]. In some of the languages, there may be only one /s/ phoneme with various realizations, while in others there are two or three sibilants. The main point of confusion has been between lamino-alveolar [s] and apico-alveolar [ṣ]; these can be called plain and retracted. It has been observed that speakers of English use both of these articulations for /s/, although the apical version is not so retracted that the acoustic difference is striking. When [ṣ] is sufficiently retracted, however, one can hear the difference fairly easily. The major conclusion that Bright reaches is that :
This point is illustrated on the next few pages by means of maps that show the languages with various contrasting sibilants. |
|
|
|
The widespread existence of this unusual contrast is a signal feature of California. Notably, it is not correlated with specific genetic groupings, and therefore must reflect diffusion within a linguistic area. |
|
These maps identify languages nearby that have similar sibilant systems — for most of these, there is one sibilant that is retracted or alternates with the non-retracted type. These languages have no close relation with most of those shown in California, but English-speaking field workers had similar problems in nailing down the contrasts.
For example, early transcriptions of Alsea generally use the letter c, which at the time stood for /ʃ/, but also sometimes also use s; later transcriptions are standardized to s, and there was no contrast. |