1Week 4b. UG and L2A:Binding domains, null subjectsCAS LX 400Second Language AcquisitionParameters• Last time we looked at the V-to-T parameter thatis responsible for the differences between French(where the adverbs cannot go between the subjectand the verb but can go between the verb and theobject) and English (where the situation is exactlyreversed.• In the studies we discussed, we found very littlethat would indicate that L2’ers ever manage to“set this parameter” in a way different from thesetting in their L1.Clustering• One reason to be suspicious of the V-to-T parameterand any purported “re-setting” of the parameter is thatthe same parameter setting (that is the movement of Vto T or the lack of such movement) is supposed to beresponsible for both of the adverb placement facts.However, what L2’ers look like they do is to retain theorder that their L1 allows and additionally allow theorder that the L2 allows—the two orders are notmutually exclusive for L2’ers like they seem to be for“L1’ers”.The null subject parameter• Adult languages differ in whether theyrequire overt subjects or not.• English does:– *Go to the movies tonight.• Italian and Spanish do not:– Vado al cinema stasera. (Italian)– Voy al cine esta noche. (Spanish)‘(I) go to the movies tonight.’The null subject parameter• There is a significant cluster of properties thatseems to go along with be a “null subject” (a.k.a.“pro drop”) language..– Subject pronouns can be omitted in tensed clauses.• (Generally are except to indicate contrast)– Expletive subjects are null. (it rains).– Subjects may be postposed. (ha telefonato Gianni)– There is no that-trace effect.•(*Who did you say that left?)– Subject-verb agreement is “rich” or uniform.White (1985, 1986)• Compared two groups of subjects learning English:– 32 native speakers of (Latin American) Spanish and2 native speakers of Italian– 37 native speakers of Québec French• Did a test of grammaticality judgments, as well as aquestion formation test:– Mary believes that Fred will call his mother.– Who does Mary believe the Fred will call?– Mary believes that Fred will call his mother.– Who does Mary believe will call his mother?2White (1985, 1986)• The results weren’t all that dramatic—the NSL1’erswill quite a bit more likely to accept a sentence with amissing subject (35% vs. 8%), but with respect tojudging overt pronoun subjects, and SV vs. VS order,both groups did about the same. It doesn’t really seemlike “the whole parameter” was transferred…?• NSL1’ers were also a lot more likely to produce a that-trace violation (71% vs. 42%), though nonNSL1’ersstill produced quite a number.Al-Kasey & Pérez-Leroux 1998• Another, more recent study looking at thepossible clustering of properties in L2A ofSpanish by English speakers.• Observations: English SSL students are knownto make errors in which they will(inappropriately) “overuse” subjects, using “toomany” subject pronouns or even fabricatingexpletives (es ‘is’, hay ‘there-are’, el ‘he’). Thisis a fairly predictable “transfer” since Englishrequires overt subjects.Al-Kasey & Pérez-Leroux 1998• Another error that English SSL students seem tomake frequently is like this: (Lee 1987, Al-Kasey& Weston 1992)…y la lee…and it-acc-fem-sing reads‘…and reads it.’• Students will quite systematically misinterpret laas being a subject (not an object which it“obviously” is to any native Spanish speaker,because it is actually marked as being an object).Al-Kasey & Pérez-Leroux 1998• A study was put together in order to look atcorrelates of the null subject parameter and to seeif they all more or less appear together or not.• Specifically Al-Kasey & Pérez-Leroux looked atplaces where English and Spanish differ withrespect to null subjects (i.e. places where a subjectpronoun is optional, places where a subjectpronoun is inappropriate, and places where aninitial pronoun isn’t a subject).Al-Kasey & Pérez-Leroux 1998• There was pretty clear evidence of transferof the subject properties of English to the IL(Spanish-to-be)—an overuse of subjects, atendency to suppose that overt subjects canbe expletive subjects. The more advancedlearners recovered, became more native-speaker-like.Al-Kasey & Pérez-Leroux 1998• The use of null subjects seemed to appear firstfor expletive (meaningless) subjects (i.e. forthings like rains), and a little bit later for silentbut meaningful subjects (like you, etc.).• There is clearly an implicational relation; if youhave null meaningful subjects, you have nullexpletive subjects.• The two properties are at least connected.3Al-Kasey & Pérez-Leroux 1998• Moreover, the levels of null subject productionachieved by the “advanced majors” werebasically the same as the levels exhibited by thenative speakers.• The correlation and the success suggested to Al-Kasey & Pérez-Leroux that we are in factlooking at connected properties, a parameterbeing set. The time lag might make us a little bituneasy, but it is a correlation.So…• The V-to-T parameter from last time seemsto be hard to “re-set”—perhaps it even can’tbe re-set.• The null subject parameter has given us lessthan clear-cut results—there seems to besome relation between the properties weattribute to the parameter, but they don’tmove directly together.Binding Theory: once more1) John saw himself.2) *Himself saw John.3) *John said Mary saw himself.4) *John said himself saw Mary.5) *John saw him.6) John said Mary saw him.7) John said he saw Mary.• Binding Theory. Principle A: Anaphors (like himself)need an “earlier” antecedent within its binding domain.Principle B: Pronouns (like him) cannot have an “earlier”antecedent within its binding domain.• Parameter: Binding domain = sentence containingBinding Theory parameter: thedomain for anaphors24) Sam believes [that Harry overestimates himself]25) Sam-wa [Harry-ga zibun-o tunet-ta to] it-ta]Sam-top Harry-nom self-acc pinch-past-that say-past‘Sam said that Harry pinched (him)self.’More advances in BT• This parameter of binding domain has beenstudied rather extensively in both theoreticallinguistics and second language acquisition.• Eventually, it was noticed that anaphors whichseem to be able to get their referent “long-distance”tend also to be monomorphemic—this isparticularly clear for languages that have