This question (or something like it) was posted on twitter some time in May and it led to a bit of back and forth, with not a lot of consensus on the right answer. I contributed a few tweets, but because I was travelling at the time it was hard to engage properly in the ongoing conversation. Moreover, Twitter isn’t a great medium, imo, for the kind of discussion the question deserves. I said I might write something about this when I was back home, and it’s taken me a bit of time to settle in and actually write up my thoughts, but here they are. Fair warning, I have a lot to get done in my last week or so of sabbatical, so I wrote this quickly and hence it isn’t very polished. It also isn’t as nuanced as it could be, but for me, for now, it’ll do. (A new attitude I’m trying out. For those who know me, not a natural way of being for me.)
It’s a deceptively simple question, and it’s not uncommon for textbooks to say that the average child has acquired most of the basics by 3 or 4 years of age. Not surprisingly, then, the OP got a few responses along the lines of “I’ve heard 3 or 4.” As well as ones that went on to say things like “but I think it’s more like 5 or 6.” Then there were a whole lot of other responses that were deemed (by someone else, not the OP) to be unhelpful. I assume that the unhelpful comment was about responses that basically said “never” or even worse “the question doesn’t make sense”. I am in the camp of the latter response, and here’s a bit of an explanation as to why. I’ll break down a few of the reasons I think the question is troublesome.
I should be clear that I don’t mean to pick on the OP: they asked what to many people seems like a perfectly reasonable question, and any points they (or others) brought up in response to answers were likely quick responses in medium that is not good for long, nuanced, reasonable exchanges. So I use them as ways to talk about why the question is problematic, not why the person is wrong. Because of this, while I use things that were said in my response, I won’t ID any of the twitter handles.
1) What do you mean by ‘their language’? To answer a question about when something happens you need to know what the something is. So what does ‘their language’ mean.
In an effort to be more specific (which is good) in response to the question of what they meant by language, the OP distinguished between ‘their idiolect’ and ‘their language’. But since any language is really just a label we give to a collection of idiolects, I am assuming they meant shared vs. non-shared things. But even that is not as unproblematic as one would like, especially given we are talking about things that have been learned. Whether one thinks that some things about language are innate or not, the things that are presumed, by some, to be innate are not the things the OP was wondering about. And there is emerging evidence that there is more variation in language knowledge in adults than we might have assumed, meaning that we would also have to define what we mean by ‘shared’ (i.e. the stuff that is part of their ‘language’). If it’s 100% agreement among adult speakers of the language, then there might be less left than we think (Dabrowska, 2018), as anyone who’s ever sat in on an undergrad (or grad for that matter) syntax class can attest. I’ve never heard a native English speaker say that “ran the dog after cat the” is grammatical, but you get much beyond basic word order and you inevitably have at least a few dissenters here and there. That’s not to say that there is not high levels of agreement on many things, but a high level of agreement is not total agreement. We could of course include things without perfect agreement, but that would require some specific pre-defined level of agreement (given a specific sample size) to include things where there is less than perfect agreement. Moreover, we often appeal to idiolects to explain (within dialect) variation. When you say ‘in my idiolect it works this way’ it means that the version of the language you have acquired works that way, that is, you are including the idiolectal in the language that has been acquired.
This may seem like an absurd thing to bring up, but we need to define what we are asking about, and even that basic step is more fraught than it would at first appear. I mean, we all know what we mean when we say the child is learning English, or French, or ISL, right? Yes, of course we do, but saying the child is learning English is different than making a claim about when she is has actually ‘acquired’ English. That entails a target that is reached, and my point is that it’s not totally clear what that target includes, so how can we decide that it has been reached. It’s kind of like asking someone how long it will take to drive to Toronto from somewhere. Do you mean to the outskirt of the region people refer to as Toronto, the actual boundary of the city proper, or to downtown? Those are 3 different places and they take three different times to reach.
2) Even if we could distinguish between their ‘language’ and their idiolect, there are still a lot of ways we could define ‘language’ and this was pointed out. Setting aside for now what we mean by ‘acquired’ (more on that later) do you care about the phonological system, the syntax, the semantics, the pragmatics, sociolinguistic variation? All of it? Many people would likely leave out the lexicon, as we know that we continue to learn new words throughout our lifetime and it feels odd to say that someone who doesn’t know every word in a language (i.e., every word used by some speaker) hasn’t yet acquired the language. (But even then, would we want to say someone who only knows 500 words knows the language? I don’t know the answer to this question, but the point is that these sorts of questions are tricky.) But if we care about the learning of culturally shared things, then things like socio-linguistic variation would seem to be especially relevant, as no one, regardless of their theoretical orientation, would suggest that these things aren’t learned. Other things fit less clearly (e.g. affix ordering). Some people thing everything is learned, others think some things aren’t. So your theory can define what you include in the set of things that are learned and so is relevant for answering the question. Here I am not taking a stand on what is versus what isn’t learned, my views on that are not relevant to the point I am making. What is relevant is that what you include in “their language” is not theory neutral, if what you want is to include only things that are clearly learned.
3) Then there is the question of what ‘acquired’ means. There was some mention of an inflection point, where, e.g., learning stops being fast and starts being slow, in order to accommodate the known ongoing changes in adult language (e.g., changes in RC processing, Wells, Christiansen, Race, Acheson, MacDonald, 2009), the idea being that the inflection point would be the time of interest. In principle, the idea of an inflection point is reasonable (at least to me), but since each form in the language likely has its own, deciding on how to create a super measure that would include them all (even within a specific domain of interest) would be arbitrary. (The extant data suggest that each individual aspect of language has its own trajectory, where by individual aspect I don’t mean things like ‘passives’ or ‘relative clauses’ or ‘tense’, I mean things at a more fine-grained level than that: e.g., verb agreement appears to be acquired on a morpheme by morpheme basis (possibly even a verb+agreement morpheme by verb+agreement morpheme basis. So you’d need an inflection point for each individual ‘thing’). I am not saying that we couldn’t create a measure of average inflection points (across aspects of the language), just that, however we do it, it will be arbitrary. Moreover, I would actually be surprised if every child showed the same ‘super-line’, that is, if an averaged inflection point looked the same for all children (it certainly doesn’t for word learning, e.g., see work on the naming insight/naming explosion). And for children in different cultures where there are different practices surrounding talk to children timing may be affected (Shneidman & Goldin-Meadow, 2012). (It’s easy enough to have different ages at which the language is considered acquired for different cultural and linguistic groups, but that has the potential to bring with it a lot of unnecessary baggage.) Additionally, what are we measuring? Production? Comprehension? Generalization? Correct generalization into all possible contexts? All of these?
Basically, while it’s possible, in principle, to figure out exactly what you mean by acquired, for what, and how you are going to measure it, it’s not at all clear to me that it’s in any way interesting, or more importantly, meaningful, to do so, because time point that comes out as a result will be an arbitrary one, that is only meaningful within the specific definitions. For these, and other reasons then, I don’t think the question makes sense.
Now back to my sabbatical to do list in a vain attempt to get one or two things I had planned to do actually crossed off before it ends.
Dabrowska, E. (2018). Experience, aptitude and individual differences in native language ultimate attainment. Cognition, 178, 222-235.
Shneidman, L. A., & Goldin-Meadow, S. (2012). Language input and acquisition in a Mayan village: how important is directed speech?. Developmental science, 15(5), 659–673. doi:10.1111/j.1467-7687.2012.01168.x
Wells, J. B., Christiansen, M. H., Race, D. S., Acheson, D. J., MacDonald, M. C. (2009). Experience and sentence processing: statistical learning and relative clause comprehension. Cognitive Psychology, 58(2), 250-271. doi: 10.1016/j.cogpsych.2008.08.002.
Hi Carla! Thanks for the detailed thoughts in response to my question. (I appreciate the good reasons not to tag me explicitly – for which thank you! – but I’m happy to out myself, so to speak.)
You might be surprised that I agree with a great deal of what you write here. I particularly agree with the need to be precise about terms. Indeed that’s exactly why I put the words “end” and “acquired” in scare quotes back in my original tweet, because they are loose terms. I agree also with your general observations about how to make them more precise.
In fact I only have one substantive point of disagreement; and happily I think this is the nub of the matter.
You write, “I am not saying that we couldn’t create a measure of average inflection points (across aspects of the language), just that, however we do it, it will be arbitrary.” I don’t see why describing inflections points must be arbitrary.
We could (in theory) describe a distribution curve of inflection points for a given idiolect. That wouldn’t be arbitrary. On the contrary, it would be an empirical fact. It would be a precise and rich answer to my original question.
We could proceed as follows. We carve up an idiolect in some sensible way (i.e. we define the various parts of a language that are to be acquired); we collect the data on the acquisition of those parts; and we collate the observed inflection points into a distribution curve of acquisition for an individual. This would be the “super-measure” you mention. (“Super-measure” is a nice phrase btw.)
And if we did the same for many individuals with the “same” L1, we could generate a super-super-measure that would answer the question for that L1. Subject to good data and sufficient computing power, we could even go further and generate a super-super-super-measure for languages in general.
One might ask, why bother? The answer is, in short, because good description is essential for science. I can’t see why we wouldn’t want to know what it is that we trying to explain when we study language acquisition (or anything else).
Thanks again for the detailed engagement!
Scott, thanks for taking the time to read the post, and reply. I still think it would be an arbitrary measure, because of all of the decisions that would need to be made along the way, including, what counts as an inflection (which the word learning literature has shown isn’t as straightforward as we once thought). Given all of the decisions that would have to be made along the way, I would worry that any measure would be lent an aura of objectivity that I don’t think would be warranted. We may just have to disagree about this.