But some is less universal than other(s).
I assume that Language Log will probably pick up the story Language universality idea tested with biology method from the BBC and, no doubt, have much wiser things to say about it than me. However, it’s awhile since I had a tale for the languages and linguistics section, and when I see Universal Grammar (UG) questioned, I’m a little curious.
UG is the idea that the brain has some kind of structure which is configured when we acquire one or more languages as children. It’s been central to the theory of grammar for about 50 years or so, and how linguists describe language (all right, phonology since that’s my background) has changed from SPE with its linear rules, to non-linear phonology (with rules), to principles and parameters, to Optimality Theory. But UG remains the underlying assumption in all of this.
The boys on Language Log don’t tend to have much regard for BBC science reporting, and since I have little or no knowledge about evolutionary biology, I can’t really assess the research from that perspective.
Early in the article I read
The results suggest that features shared across language families evolved independently in each lineage.
I’m not exactly certain how I should interpret this. For one thing, although it’s possible to identify the members of the Indo-European language family, I’m not aware of there being much certainty with respect to the interrelationships among the different sub-families. Each branch seems to be largely independent of every other branch.1 I assume that that’s because Indo-European was overlaid on different substrates.
The authors say cultural evolution, not the brain, drives language development.
This seems to be straying into sociolinguistics. I don’t disagree that some language change is driven by the society in which a language is spoken. An obvious one is, for example, prestige-driven changes, whether these are overt or covert. There’s probably a large amount of language change which has been purely ephemeral. Some changes are à la mode one moment and passé the next.2
“We show that each of these language families evolves according to its own set of rules, not according to a universal set of rules,” Dr Dunn explained.
The set of rules is particular to a language, but is drawn from UG. In other words, this is looking at language change from the wrong direction. When a child acquires a grammar, it formulates its own version of that grammar. By and large, it agrees with everyone else’s, but there are always parts that are different. For example, which and witch are homophones in my idiolect, but I’ve encountered people for whom they’re a minimal pair, and I know my parents distinguish the two (although the initial semi-vowel of which does not seem especially distinctive in their speech). Why didn’t I acquire this feature? I don’t know, but its absence is an aspect of language change in my own speech. Possibly, my brain filtered out the marked segment. In addition, I can see no cultural dimension to this change.
Whatever I was doing, it had nothing to do with UG, which is a mental framework and not some guiding principle for how languages ought to develop.
I am also a little sceptical about the particular focus of the paper, which was adpositions and their relationship to clause order. As a general principle, right-headed languages such as English have prepositions and VO clause structure, and left-headed languages have postpositions and OV clause structure. There are exceptions because in English, for instance, adjectives are pre-head modifiers (e.g. the fat cat) unless they’re modified themselves (e.g. the cat fat on cream). Here the BBC article is a little hazy on the exact details.
Steven Pinker, whose book The Language Instinct I’m reading at the moment, is left to say, “Er, well, maybe.”
Overall, paint me sceptical about this. Some parts appear to be nothing more than things that we already knew (i.e., the sociolinguistics of language change), while other parts seem (deliberate italics what with the English verb lacking the sceptical mood) to misunderstand grammar (namely, UG).
Since I wrote that, I’ve found that Language Log does have an entry about this story (Word-order “universals” are lineage-specific?), which points readers to a summary of the original article here. I remain sceptical on two counts. One is that I can’t recall seeing any claims about the setting of the VO or OV parameter in a language unquestionably determining the settings for other parts of phrase structure, but I may be quite wrong about this. The other is that the tendency to harmonise the headedness of phrases would not seem to have anything to do with UG per se because, I would hypothesise, UG is platform neutral. In other words, headedness may be a consequence of processing efficiency or the inclination of the mind to observe patterns or even a little of both. Strictly idle speculation on my part, of course.
It’s also worth noting that languages have their quirks which cannot be related to UG. For example, the English vowel system appears to have been going around in circles since some time in the Middle English period. Low and mid vowels are raised; high vowels are diphthongised. The big change to the vowel system was the well-known Great Vowel Shift, and English seems to have got onto this vocalic conveyor belt as a result. This is not a UG thing, nor determined by UG, but the resulting structures can be accommodated by UG. Why the English vowel system apparently behaves like this, I don’t know, but I think it stands outside of UG myself.
1. It’s obvious, though, that certain groups of families do have a shared inheritance. The obvious one is the division between the centum and satem languages, the latter sharing at least one round of palatalisation. However, the satem languages were presumably a collection of dialects at the time, which would eventually become Indic, Iranian, Armenian and other language groups.
2. Examples from English (if these are not myths) are the treatment of high-mid and low-mid vowels in London English in the 17th (?) century, and the reappearance of the velar nasal in the -ing ending which, allegedly, was a consequence of the vulgar mob aping the quality who then (so I’ve read somewhere) are supposed to have reverted to the -in’ form.
(Cross-posted from Green Bamboo LJ.)