"Around seven thousand languages are spoken worldwide—a seemingly bewildering array of sounds, words, and rules. Yet for many linguists, this Babylonian confusion exists only on the surface. Beneath it, they seek structures shared by most, if not all, languages. Candidates for such universals include rules of word order or the grammatical ways in which a sentence distinguishes between the agents, objects, and those affected by an action. Such shared patterns can offer clues regarding the cognitive or communicative constraints that shape the evolution of all human languages.
The "Konstanz Universals Archive"—a database at the University of Konstanz—now lists over 2,000 universals; or, more precisely, phonological and grammatical structures that linguists have classified as such. Often, however, these classifications were based on limited empirical evidence. This raises the question of whether the label of "universal" was always justified. A research team led by Annemarie Verkerk (Professor at Saarland University) and Russell D. Gray (from the Max Planck Institute for Evolutionary Anthropology in Leipzig) has now used statistical linguistic methods to examine 191 of these patterns and determine whether they are truly globally distributed.
The data for this study came from "Grambank," a database maintained by the Max Planck Institute for Evolutionary Anthropology that stores grammatical features of nearly 2,500 languages and dialects. German, however, is not included; nevertheless, many of its structural characteristics are covered by closely related and grammatically very similar languages—such as Dutch and Frisian—that are featured in the database.
Even Annemarie Verkerk cannot answer why German—one of the world's largest and most thoroughly studied languages, and the language of the country where Grambank is based—is missing. Several other major languages are also absent, including Spanish and the Indian languages Bengali and Gujarati. However, according to the linguists, these gaps posed no problem for their study on universals.
The researchers examined 1,700 of the languages documented in Grambank to see whether the selected grammatical structures occurred there. The result: No fewer than 60 of the 191 hypothesized features proved to be genuine universals.
The study focused not on isolated features, but on grammatical rules that correlate with one another. Word order serves as an example: in languages like German, where the adjective precedes the noun (*dunkle Wälder* / dark forests), articles (*die Häuser* / the houses) and numerals (*drei Häuser* / three houses) are placed similarly, and prepositions (*auf dem Berg* / on the mountain) predominate over postpositions.
Furthermore, constructions expressing possession or association—such as genitives and relative clauses—usually follow the head noun (*Haus des Vaters* / house of the father; *Haus, das dem Vater gehört* / house that belongs to the father).
These same patterns are also found, for instance, in Mandarin Chinese—a language that is unrelated and geographically distant—which is a crucial prerequisite for true universality.
The mirror image of this pattern is found in languages where word orders are exactly the reverse. Another example of a universal concerns the relative position of the object and the verb. If the object follows the verb (*sie sieht den Baum* / she sees the tree), prepositions predominate. If the reverse is true—as in Japanese or Basque—postpositions are the norm. However, some of the grammatical patterns used in the study of universals are based on the simplification of mixed conditions. In German and structurally similar languages, for example, the verb precedes the object in main clauses but follows it in subordinate clauses (*als er ein Haus baute* / when he built a house); the genitive can also precede the head noun (*Vaters Haus* / father's house), and prepositions do exist (*gemäß der Regel* / according to the rule) as a postposition (in accordance with the rule). It is unclear what impact smoothing out such variations has on the study results.
Many linguists believe that the fact that different word-order rules within a language often align in similar ways stems from cognitive factors. In doing so, speakers unconsciously impart a consistent information structure to their languages. If one rule within a language changes, the others often follow suit.
One challenge in identifying universals lies in ruling out similarities that are merely the result of language relatedness or arose through borrowing from one language to another. For this reason, linguists have previously often selected languages for their studies that are geographically far apart are separated. However, this restriction to isolated samples reduced statistical power and also ignored history: it failed to account for the possibility that some of these languages might have influenced one another in the past or even shared common ancestors.
To circumvent these methodological pitfalls, Verkerk and her team analyzed the Grambank data using statistical models that take such factors into account. "Using our methods, we were able to trace how a language evolved historically and determine its geographical relationship to other languages," says Verkerk. According to the linguist, this approach made it possible to distinguish genuine universals from apparent ones and achieved a higher level of statistical precision than previous studies.
Given that only 60 of the 191 proposed universals proved to be statistically supported, the question arises: is the glass two-thirds empty or one-third full? Co-lead researcher Russell Gray believes the latter is the case. "Patterns are emerging here that develop repeatedly. They demonstrate that, driven by shared cognitive and communicative conditions, languages tend toward a limited number of preferred grammatical solutions."” [1]
1 Hinter dem babylonischen Sprachgewirr: Ein Forscherteam untersucht, was allen Sprachen gemeinsam ist. Frankfurter Allgemeine Zeitung; Frankfurt. 10 June 2026: N4. WOLFGANG KRISCHKE
Komentarų nėra:
Rašyti komentarą