Impact of Cross-Linguistic Research in Studying Specific Language Impairment Behind the Science
Behind the Science  |   November 01, 2013
Impact of Cross-Linguistic Research in Studying Specific Language Impairment
Author Affiliations & Notes
  • Laurence B. Leonard
    Purdue University
  • The content of this page is based on selected clips from a video interview conducted at the ASHA Convention.
    The content of this page is based on selected clips from a video interview conducted at the ASHA Convention.×
Article Information
Cultural & Linguistic Diversity / Research Issues, Methods & Evidence-Based Practice / Language Disorders / Specific Language Impairment / Clinical Practice Research / Overview, Trends and Environment
Behind the Science   |   November 01, 2013
Impact of Cross-Linguistic Research in Studying Specific Language Impairment
CREd Library, November 2013, doi:10.1044/cred-ote-bts-001
CREd Library, November 2013, doi:10.1044/cred-ote-bts-001

In recent, I can almost say recent decades now, I've been very active in cross-linguistic research. Initially I was focused on English-speaking children, and we really had a handle on what these children are like. There are some individual differences, of course.

Somewhere along the line, I read something about typical language development across languages, and I noticed they're describing these children who are acquiring other languages as being quite different from ours. So, I started wondering, what would language-disordered children look like? Children who basically had the same diagnosis, but who were acquiring a very different kind of language.

I started eventually studying children in Italian with a diagnosis of specific language impairment, only to see that their symptoms were really quite different. There were a few universals, like most of these kids were slower in acquiring words. Most of these kids were slower in putting words together into two-word combinations. But beyond a few of those more or less universal characteristics, their grammatical patterns for instance, in terms of what was especially problematic, were quite different from those in English.

That was very eye-opening to me. Things that I thought were "abstract" and maybe beyond the reach of children with language impairment, especially kids at the preschool level, turned out to be a snap for Italian-speaking children with language impairment. Because there is a lot of redundancy and transparency in the language, very overt agreement between adjectives and nouns and subjects and verbs that we just don't have in English. And the information at the ends of words, which can be problematic for English-speakers, because these are consonants like t's and d's and s's and z's, are vowels in Italian. So what proved to be difficult for these kids included inserting little function words, but not producing noun or verb endings, for instance.

When studying some of the other Germanic languages, after moving from Romance languages to Germanic languages, I'm realizing now that word order is problematic for children with language impairment in other languages -- it never occurred to us in English, because English is such a strict subject-verb-object kind of language, where we don't deviate from that except in a few specific instances. So word order is just not problematic for children in English, but it sure is for kids in other languages.

So, that made me have to dig further. It's not going to be something simple like, grammatical agreement is difficult, or the notion of plural is difficult, or the notion of past-tense is difficult. It's not at all simple. There's a heavy interaction between what's transparent in the language, and what's peculiar in the language, and what's a grammatical function. It's been pretty exciting from that standpoint.

How have the cross-linguistic studies changed your approach to investigating SLI?

What you need to do now is try to find what could be responsible. Of course, we're always shooting for the "single factor." It's rarely or never quite that simple. But you're hoping to find some kind of overriding factor that might account for very different symptoms as a function of the kind of language being acquired.

Just to give you an example, when I looked at all these differences after making an observation that doesn't sound very profound at all, that is: Whatever features, we're talking about grammatical features now here, happen to be problematic for young typically developing children acquiring that language. For instance in English, in our language acquisition textbooks, we often hear about how young children are telegraphic in their speech. They produce these two and three-word utterances without words like "the" and "of" and "is." And sometimes they leave off "ed" and "s" at the end, so it's very telegraphic. That metaphor works very well. We think that's the way all kids are going to be. Then we find children in some of these languages that have very rich grammar, where all words have to have a grammatical ending. A kid never hears a word without a grammatical ending. And sure enough, they never leave off grammatical endings, no matter how limited they are, because they never hear one and it never occurs to them. How can I account for these very different kinds of symptoms, and yet these kids do have a language problem very clearly.

That's when we started to think about -- I'll give you another example. Some kids will say things like "him running" or "her running." It sounds like, where did they get that? It's very ungrammatical. Or "him play basketball." Why don't they say "He plays basketball", "He played basketball"? Where is this "him" and "her" coming from? Well, as we started thinking about it, we realized that in the input, kids do hear things like "I saw him play basketball yesterday." Or, "I saw him running." So you can get errors of that type, and then think, if kids -- and we do know they have a wide variety of problems -- if kids are not understanding the more complex sentences, what they're doing is inappropriately extracting things like "him play basketball" because they don't understand that you can only say "him play basketball" if you have something earlier in the sentence that allows for it. If they don't understand that larger grammatical structure, they might inappropriately extract things.

What's interesting is, in let's say a language like German, you have word order errors that are very un-English like. For instance, a child might alternate between -- I always use the example -- "Christina drinks coffee" (but the German equivalent), they'll alternate between that correct form and an incorrect version like "Christina coffee drink" where drink will have the infinitive ending. How do you explain that? Well, as a matter of fact, when you ask questions like "Can Christina drink coffee?" in a language like German, you would actually ask it correctly as, "Can Christina coffee drink?" So if kids, again, don't understand that the use of drink in a sentence is going to depend on something earlier on in a question like "can" -- they might inappropriately extract something like "Christina coffee drink." and alternate that with "Christina drinks coffee." because they also hear that in simple sentences.

I guess the point is that if you make a single assumption, that there are certain larger syntactic structures these kids have a hard time deciphering, you can get a very varied range of errors that look, at first glance, like these are totally different things going on if a kid is in Germany versus the US or in Italy or in Israel or in Hong Kong or Hungary or Finland -- because we are studying kids in all those countries -- that these are very different errors. But sometimes you can account for what seem to be different errors by making the same assumption.

But we never would have thought about that if we weren't forced into looking at very different errors according to the kind of language the kid is acquiring.

How can cross-linguistic research benefit the "big picture" of clinical practice research in CSD?

We never could have done that with English alone. And someone in a different country could never do it by looking at their language alone. That's absolutely critical.

But there's another benefit as well, and that is the more we understand -- for instance some of the collaborative work I've done with colleagues who were studying monolingual Spanish speaking children with specific language impairment, and we would sometimes compare with English and so on -- it turns out the more you know about those languages, the more you're in a position to study something that's really important in the US these days, and that is bilingual children with specific language impairment. Bilingual Spanish-English children with specific language impairment are a growingly important (if there is such a word) participant population these days because there are a lot of bilingual kids, and some of them will have language impairment, just like a certain number of monolingual children will have impairment.

So the cross-linguistic research is also important for understanding the basic condition, but the more we can catalog details about different languages, the more we're in a position to also help kids who have more than one language. And how those languages might interact. Because we know how each language works separately, first of all -- if we didn't know that, we'd be stuck with bilingual. I think for the longest time, a lot of the work in the bilingual area emphasized important things like cultural values and things of that sort. But, this kid has this problem now. What are you going to do? You can respect the child's culture, but now what are you going to do -- when you have this language and this language and the child is using both of them? You really need a lot of those empirical facts as well. I think the cross-linguistic work has really helped to bring some of those facts.