Language and Logic
In the normal course of events, children manifest remarkable linguistic competence in just a few years. Research on child language has shown that, by four or five, children around the globe produce and can understand novel sentences, they can judge that certain strings of words are true or false, and so on. Linguistic research has revealed that all human languages (e.g. English, Basque, Swahili, etc.) have certain linguistic properties in common. It has been proposed by advocates of Universal Grammar that many of the shared properties of natural languages are innately specified as part of the human biological blueprint for language development; this is why human children learn languages so effortlessly, and so rapidly. Advocates of this theory contend that experience dramatically underdetermines the linguistic competence children achieve, even given optimistic assumptions about children's capacities to extract information and form generalizations on the basis of observable regularities in the input. In light of these observations, a central issue in the 'nature vs. nurture' debate about child language is the extent to which principles specific to language 'grow' in children, as part of the human biology, and the extent to which knowledge is acquired by children using the same general-purpose cognitive mechanisms used to extract information and form generalizations about 'non-linguistic' things, such as how to add and subtract.
This project investigates facts about logic and language that remain hidden from learners, yet facts that every learner figures out, usually early in the course of language development. The theory of Universal Grammar supposes that young children solve these mysteries by drawing upon a priori knowledge of universal linguistic principles and parameters. One set of mysteries concerns children's emerging knowledge of the interpretation of natural language disjunction (e.g., English "or", Chinese "huozhe", Japanese "ka"). It reasonable to suppose that basic meaning of disjunction in natural language is the same as in classical logic, namely 'inclusive-or' (such that A or B is true if either A or B is true, and if both A and B are true). The research question becomes: how do children figure out the basic meaning of disjunction?
Another project investigates a semantic property: 'downward entailment.' Downward entailment ties together the interpretation of several disparate phenomena such as the where words like any and ever are licensed, and the interpretation of disjunction. We have found that by 4-years old, children understand the concept of downward entailment. In future research this line of exploration will be extended to new constructions, using tried-and-true behavioural measures that have been successfully applied in previous research (e.g. the Truth Value Judgment Task). We must begin to study children younger than four, however. Research on younger children is important, because this will compress the time frame for 'learning', thereby adding to the arsenal of empirical data that can be used to evaluate the competing theories of language acquisition. Finally, we will begin to explore the brain indices of linguistic phenomena involving downward entailment, using magnetoencephalography (MEG). This technique will be used in children and adults, for example, to see how the brain responses to the disjunction operator "or" in minimal pairs of sentences.
Below are some individual projects currently being undertaken in collaboration with Stephen Crain and Rosalind Thornton:
The word 'every' in child language
Anna Notley, Britta Jensen, Francesco Ursini
This research investigates whether young children know what a sentence like 'Give every bear a bowl' means. If presented with a row of 5 bears and a stack of 6 bowls, adults would most likely hand out one bowl to each bear and discard or ignore the extra bowl. Some theories of children's understanding of the word 'every' predict that, unlike adults, children might think they need to hand out all the bowls. That is, they might think the word 'every' applies to bowls rather than bears, or they might think it applies to both bears and bowls.
We show in experimental play sessions with children aged 2-3 that when the conversational context is natural, children have no problem restricting the word 'every' to the subject noun phrase only, in this case the bears only. They consistently show us that the extra objects (e.g., bowls) are not relevant for them. It has been suggested that the way in which quantificational determiners (like 'every', 'some' and 'few') are interpreted may be universal across human languages. If this is true then when children first begin to understand the word 'every', they should access the universal interpretation. Our data support this.
The words 'before' and 'or' in child language
Anna Notley
The word 'before' interacts in an interesting way with the word 'or' when the two words are in the same sentence. In a sentence like 'The principal arrived at school before the teachers or the students,' we understand that the principal arrived at school before both the teachers and the students. However, in a very similar sentence with the word 'after' such as 'The principal arrived at school after the teachers or the students' we do not understand that the principal arrived at school after both the teachers and the students. Rather, the principal must have arrived after one group only, although we are not sure which.
This property of 'before', as opposed to 'after', is due to certain logical inference rules. We are interested in exploring which of these logical rules may be part of our innate cognitive make-up. This research will test English speaking and Mandarin Chinese speaking children to see whether, in the context of a running race, they understand a sentence like 'The dog arrived at the finish line before Tigger or the bunny rabbit' like adults do, or whether they think it means that the dog arrived before only one of the other participants in the race.
Children's Understanding of "mei-ge...huozhe..." in Chinese
Yi (Esther) Su
The Chinese word 'huozhe' (English 'or') displays an interesting asymmetry when it appears together with 'mei-ge' (English 'every'). When 'huozhe' appears in the subject phrase of mei-ge, as in 每个点了寿司或者意大利面食的男孩都得到了碟子。(English: 'Every boy who ordered sushi or pasta received a plate'), we understand that every boy who ordered sushi received a plate and every boy who ordered pasta received a plate. However, in a very similar sentence with 'huozhe' in the predicate phrase of 'mei-ge', such as 每个男孩都点了寿司或者意大利面食。(English: 'Every boy ordered sushi or pasta'), we don't understand the sentence to mean that every boy ordered sushi and every boy ordered pasta. To describe a context in which every boy ordered both sushi and pasta, adult speakers will use the word 'he' (English 'and').
Using the Truth Value Judgment Task, this study aims to investigate the awareness of this asymmetry with 'huozhe' and 'mei-ge' in 2-to-6-year-old Mandarin-speaking children. In a typical story, four aliens come to a restaurant on earth. Two of the aliens order pizzas and receive ice-creams from the waitress, whereas the other two aliens order hamburgers and receive strawberries. Then a puppet (Kermit the frog) will describe what happened with a sentence like 每个点了比萨饼或者汉堡的外星人都得到了冰激凌（English: 'Every alien who ordered a pizza or a hamburger received an ice-cream'). Chinese-speaking adults reject this description of the story, because the two aliens who ordered hamburgers hadn't received ice-creams. We are interested in finding out whether children, like adults, reject Kermit's description of the story, or whether they think it's true. If children think the sentence means every alien who ordered pizza got an ice cream or every alien who ordered a hamburger got an ice cream, for example, then the sentence would be true, and children should agree with Kermit's statement.
Is there only one meaning of 'or' in child language?
Nobu Akagi
Previous studies have found that children understand the logical word 'or', as in the phrase 'an apple or an orange' to mean 'either an apple or an orange, or possibly both.' However, we don't know whether children permit the same range of possible meanings (A or B or possibly both) when 'or' links two sentences, as in the sentence in (1), below.
(1) John ate an apple or he (John) drank coffee.
As adults, we interpret (1) to mean "either John ate an apple or he drank coffee," and we are less likely to think John did both. We want to find out whether or not children accept (1) as a good description of a context in which John eats an apple AND drinks coffee. If, unlike adults, children accept (1) in a context where John both eats and drinks, then children initially display knowledge of 'or' which differs from adult usage. If they do, we need to understand how children come to constrain their interpretation of 'or' when it is used to link two sentences. This project hopes to contribute to our understanding of how children's knowledge of language interacts with other cognitive knowledge.
Children's knowledge of 'anything'
Britta Jensen
An interesting pattern is exhibited by words like 'anything'. For example: 'anything' can follow 'before' but not 'after' and 'anything' can follow 'nobody' but not 'everybody'. Do young children know which sentences permit or prohibit words like 'anything'? This research investigates whether two- to three-year-old children appreciate the constrained distribution of 'anything' in sentences with the universal quantifier ('every' or 'all'). In sentences with 'every', adult speakers permit 'anything' to be part of the subject but prohibit 'anything' from being in the predicate. This is shown below
(1) Every bear who put anything away got a snack. [well-formed]
(2) Every bear who put away some blocks got anything to eat. [ill-formed]
Using the Truth Value Judgment Task and an Elicited Production task, we aim to find out whether young children show mastery of this asymmetry about 'anything' in the context of 'every'.
Scope Interpretation in Chinese
Peng Zhou
Consider the sentence 'Every horse didn't jump over the fence'. In English, this sentence is ambiguous. It can mean either that (a) none of the horses jumped over the fence, or (b) not all of them did. This difference in interpretation is referred to as a “scope phenomenon.” In contrast to English, Mandarin Chinese exhibits scope rigidity, in the sense that scope is determined exclusively by the surface order of quantificational expressions. In Mandarin, the sentence 'Every horse didn't jump over the fence' is 'Mei-pi ma dou meiyou tiaoguo liba' (每匹马都没有跳过篱笆). This sentence of Mandarin can only mean that none of the horses jumped over the fence; the (a) reading in English. We are investigating the possibility that Mandarin-speaking children differ from adult speakers in scope assignments. That is, we are exploring the possibility that Mandarin-speaking children initially find sentences like 'Mei-pi ma dou meiyou tiaoguo liba'（每匹马都没有跳过篱笆）to be ambiguous, just as it is in English. If so, we are interested in finding out how Mandarin-speaking children converge on the same grammar as adult speakers, without ambiguity.