Final r-cluster metathesis in one child's French

My favourite 4-year-old is doing something very interesting these days with final consonant clusters in his French. Many word-final consonant clusters starting with R get metathesised: parle (speaks) becomes [palʀ] (yet parler "to speak" remains [paʀle]), tourne (turn) becomes [tunʀ], herbe (grass) becomes [ebʀ], ferme (close) becomes [femʀ]. On the other hand, "porte" (door) remains [pɔʀt]; regarde (look!) [ʀəgaʀd]; "force" (strength) [fɔʀs]; "mars" (March) [maʀs], "parc" (park) [paʀk]. Presumably the phenomenon is related to sonority: {l, n, m, b} metathesise, {t, d, s, k} do not. But French allows word-final consonant clusters with falling or rising sonority, and he has no trouble with words like "monstre" (monster) [mõstʀ]. Any idea if this is typical in French first language acquisition?

Nothing of the sort happens in his English or his Arabic. Then again, his English is non-rhotic anyway for some reason, and in Arabic he pronounces /r/ as [ʕ]; French is the only one of his languages where he's got the pronunciation of rhotics more or less sorted.

Too strong to get out

At four, my nephew speaks English (his dominant language) very well. He still shows some interesting divergences from the standard of those around him, though. Some are influenced by German (a close second): he uses "mine" as a determiner in English (like German "mein") rather than "my", saying things like "mine house". Others seem to result from language-internal overgeneralization, as when he said:

• If I push the Lego box then the carpet will destroy. [intended meaning: be destroyed]

Presumably, he's interpreted "destroy" as a labile verb, like "open" or "burn".

At first blush, I thought the following sentence was another example of overgeneralization:

• I'm too strong to get out, so you can't. [intended meaning: I'm too strong for anyone to get me out]

However, reflection suggests that this ought to be perfectly grammatical in English, since "get out" is already labile. "This stump is too heavy to pull out" works fine, so why not "I'm too strong to get out"? Yet, for me at least, the clause immediately receives a pragmatically absurd interpretation with "I" as the subject of "get out", and the obviously intended interpretation is barely accessible even when I've consciously concluded that it should be grammatically acceptable.

In terms of the classic Chomskyan analysis of control, the two interpretations correspond to different unpronounced pronouns PRO:

1. I_i'm too strong [PRO_i to get out]
2. I_i'm too strong [PRO_arb to get PRO_i out]

A lot of linguists really dislike the idea of an unpronounced pronoun. Whatever its psychological merits, though, this analysis has the advantage of suggesting why the first interpretation comes more easily than the second here: it only involves one empty pronoun, whereas the desired interpretation needs two. So if anything is going wrong in this sentence, it's not so much the syntax as the pragmatics: an adult speaker might be more aware that listeners could have trouble processing a clause of this form, and avoid it in favour of something less ambiguous. That would need empirical checking though.

Can two kids change Algerian Arabic? (Probably not, but let's see.)

In central Algerian Arabic, feminine nouns are usually marked by a suffix -a, which becomes -ət when possessed. Pronominal possessors are indicated by suffixes, eg -i "my", -u "his". The lax vowel ə cannot occur in open syllables; when the suffix starts with a vowel, this is resolved by dropping it. If doing so would result in a three-consonant cluster, then, in certain cases, the latter is broken up by inserting a new schwa after the first consonant in the cluster, and geminating that consonant: thus jəfn-a جفنة "big bowl" becomes jəffən-t-i جفّنتي "my big bowl". I've been trying to figure out when exactly this happens in the dialect of Dellys, and finding a good deal of variation, especially in the treatment of sonorants: some people (especially but not exclusively the older ones) say səlʕ-t-i سلْعتي "my goods", leaving the cluster intact, while others say səlləʕ-t-i سلّعتي. I was surprised, however, to find two children, 8 and 10-year-old siblings, using a strategy not, as far as I know, used by adults for nouns at all: changing the problematic ə into a. This was confirmed not just by elicitation (zənq-at-i زنقاتي "my alley", ʕənb-at-i عنباتي "my grape", səlʕ-at-i سلعاتي "my goods") but also by sentences produced; thus for bəlɣ-a بلغة "pair of flip-flops":

ənta ʕəndək bəlɣ-at-ək w ana ʕəndi bəlɣ-at-i انتا عندك بلغاتك وانا عندي بلغاتي
"You have your flip-flops and I have my flip-flops."

and for xədm-a خدماتك "work", completely unprompted:

kəmmli xədm-at-ək كمّلي خدماتك
"Finish your work."

which his older brother actually corrected to kəmmli xəddəm-t-ək كمّلي خدّمتك.

Adults' speech furnishes one plausible model for this strategy - not in nouns but in participles. The active feminine participle takes direct object pronoun suffixes, identical to the genitive ones except in the 1st person singular. In such forms, -ət becomes -at before a vowel, rather than dropping the ə: šayf-a شايفة "having seen (f.)", šayf-at-u شايفاتهُ "having seen him (f. subject)". But its extension to nouns is something quite new; neither their parents nor their elder brother nor any adult I've met use such forms.

Most probably, the next time I go to Dellys I'll find these two children using the normal forms and denying they ever spoke this way. Even now, they already use the normal form for body parts which almost always occur possessed: rəqb-a رقبة "neck" becomes rəqqəb-t-i رقّبتي "my neck". But what if this innovation instead spreads among their peers? Most likely it won't: there seems to be little evidence for children initiating language change, notwithstanding the idea's widespread adoption by generative historical linguists, and adults' innovations are much more likely to be maintained or copied (cf. Luraghi 2013, Foulkes and Vihman fc; for a potential counterexample, see Moyna 2009). For that very reason, however, it will be worth keeping an eye on them; potential counterexamples are always interesting.

Child language acquisition and constructions

A memorable line from a talk by Ewa Dabrowska that I went to recently:

"It is generally agreed that the representations assumed by generative theories cannot be learned from the input. For generative linguists, this fact is a fundamental premise of arguments for the innateness of at least some aspects of these representations: since they cannot have been learned from the input, they must be available a priori. An alternative conclusion, of course, is that we need a better theory - one that does not assume representations that are unlearnable."

Her answer is construction grammar: kids first learn individual low-level constructions like "What's ___ doing?" as unanalysed units, and only later come up with higher-level schemas of which these constructions are special cases (the next stage in this case would be "What's ___ ___ing?") Judging from the evidence she presented, showing that the vast majority of a 3 year old's utterances could be accounted for solely on the basis of simple substitutions within sentences they are known to have already heard, "children's [linguistic] creativity seems to involve superimposing and juxtaposing memorised chunks." This view of language more or less inverts the usual grammarian's perspective: the most general rules are developed only after specific cases have been learned, and the specific cases presumably continue to be stored independently. It strikes me as a rather promising way of thinking about historical syntax.

Kant's Sparrow and the Wolf Girl

I remember coming back to Algeria after a year or so in America at the age of six. I had completely forgotten the Arabic I had known, and relearning it was an incredibly difficult process that took years, made no easier by my frequent preference for books over playmates. Over the past seven months, I've found learning Siwi and Korandje far, far easier than learning Arabic was then, and I'm pretty sure I speak both of them, if not fluently, at least far better than I spoke Arabic after my first four months back. Yet the nativist theory of language acquisition that I remember from my linguistics courses says that kids should learn languages much more easily than adults. I don't expect anybody to pick theories based on anecdotal evidence from my childhood memories, but this has made me wonder again whether kids usually learning languages faster and better is due to a pre-programmed critical period for language learning, or simply to the big difference between the social contexts of adults and children. Coincidentally (being back in London), I found two works vaguely relevant to that question this weekend; neither offers an answer, but they are interesting background.

Kant's Sparrow confirms a claim originally reported by Kant - that sparrows brought up by canaries learn to sing like canaries. Apparently, they do - but not completely. Not only do their canary songs feature a detectable accent (they differ in several ways, notably in repeating the same syllable fewer times in a row), but their repertoire includes some song types ("two-voice syllables") which they rarely or never heard from the canaries raising them, and which the author attributes to sparrows' innate repertoire (3.3.2.6.) In other words, sparrow song, like human communication, combines innate and learned (arbitrary, if you like) elements.

Wolf Child and Human Child, by Arnold Gesell, is a short, not very helpful work on a very interesting case, apparently described more fully in Diary of the Wolf Children of Midnapore, by Rev. J. A. L. Singh - two children, later named Kamala and Amala, who were adopted into a wolf family, and raised for years alongside the mother wolf's own cubs. In 1920, in response to locals' reports of a "man-ghost", a party of men dug into the wolf's den, killed the mother wolf when it tried to fight back, and brought the two children back to be taken to an orphanage (and the two wolf cubs they lived with to be sold at a fair.) Kamala was about eight, and Amala substantially younger; however, Amala died only a year later Unsurprisingly, Kamala found language rather difficult to acquire; even without the wolf factor, I imagine losing your entire family and then your entire step-family before the age of nine might have a negative effect. At any rate, apparently, she spoke her first word two years after being captured, and her first two-word sentence after three and a half years. For later years the information gets a lot sparser, but it is claimed that by the time the poor kid died (from illness) nine years later, at the estimated age of seventeen, she "talked freely with full sense of words used." The report that after several years "her formerly rigid countenance took on more expression" suggests a similar gradual development in her body language. However, while at eight years old she knew little or nothing of how humans communicate, she seems to have learned at least some wolf methods - for months at the orphanage, she would howl every night, at 10 pm, 1 am, and 3 am, and when approached by someone she did not trust she would show her teeth. Unfortunately, the lack of detail makes it hard to say what this says about first language acquisition - how well could she really speak before she died? Perhaps Rev. Singh's diary offers some quotes.

NB: see comments; apparently there is serious doubt about the veracity of this account. Looks like I should have Googled first.. The original diary also turns out to be online.