Week 11.Week 11.Maturation, passives, A-Maturation, passives, A-

chains and phaseschains and phases

GRS LX 700GRS LX 700Language Language

Acquisition andAcquisition andLinguistic Linguistic

TheoryTheory

Continuity or Continuity or Maturation?Maturation?

Pretty well accepted that there Pretty well accepted that there is is somethingsomething “built-in” “built-in” concerning the acquisition of concerning the acquisition of language (UG).language (UG).

A limiting version of this is A limiting version of this is the the Continuity HypothesisContinuity Hypothesis (Pinker 1984) (or (Pinker 1984) (or RigidityRigidity) ) which says that what’s built in which says that what’s built in is there from the beginning and is there from the beginning and doesn’t change.doesn’t change.

The situationThe situation

Quite a bit of evidence shows that Quite a bit of evidence shows that kids know a lot about the kids know a lot about the principles of UG from as early as principles of UG from as early as they can be tested.they can be tested.

Yet, languages Yet, languages dodo differ from one differ from one another—kids end up speaking another—kids end up speaking different languages depending on different languages depending on the language in the environment, the language in the environment, so they do so they do learnlearn somethingsomething..

The situationThe situation

So there are in principle two So there are in principle two dimensions of development:dimensions of development: learning language-particular learning language-particular propertiesproperties

development of the grammar itselfdevelopment of the grammar itself

Grammar development is what has Grammar development is what has been argued (poverty of the been argued (poverty of the stimulus) stimulus) notnot to be learnable by to be learnable by experience. Thus, it must be in experience. Thus, it must be in some way genetically provided.some way genetically provided.

The situationThe situation

Being genetically specified does Being genetically specified does notnot mean “present from the outset” mean “present from the outset” however. Ample evidence from other however. Ample evidence from other biological systems of this.biological systems of this.

Pretty much the only conclusion Pretty much the only conclusion available to deal with time delay available to deal with time delay of innately specified aspects of of innately specified aspects of grammar is that parts of the grammar is that parts of the grammar grammar maturesmatures..

What if we don’t like What if we don’t like maturation as an maturation as an

explanation?explanation? Two options:Two options:

Grammar doesn’t mature in a Grammar doesn’t mature in a biological sense; biological sense; it is it is learnedlearned. But . But we don’t believe that, because we we don’t believe that, because we have good reasons to think that it’s have good reasons to think that it’s just not possible.just not possible.

Grammar doesn’t mature in a Grammar doesn’t mature in a biological sense; biological sense; it is it is there from there from the outsetthe outset in its totalityin its totality. . (“Continuity”, “Rigidity”)(“Continuity”, “Rigidity”)

Neither option seems very good.Neither option seems very good.

RigidityRigidity is hard to is hard to justifyjustify

Kids don’t seem to have identical Kids don’t seem to have identical linguistic properties as adults. How linguistic properties as adults. How can we explain this without some can we explain this without some difference in the system?difference in the system?

Why do kids take so long to reach Why do kids take so long to reach adult-like competence?adult-like competence? If the data is If the data is available, why don’t kids use it available, why don’t kids use it immediately? If the learning mechanism immediately? If the learning mechanism changes, how does changes, how does itit change? change?

How far back does Rigidity go? How far back does Rigidity go? One One would suspect that “fertilization of would suspect that “fertilization of the egg” is the egg” is tootoo far… far…

The way things seem to The way things seem to bebe

We have evidence that We have evidence that kids kids dodo know know quite a bit of what we posit to be in quite a bit of what we posit to be in UGUG and very early, often as early as and very early, often as early as we can test it.we can test it.

We have evidence that We have evidence that in certain in certain areas kids’ grammars differ from areas kids’ grammars differ from adultsadults. We also have in some of these . We also have in some of these cases evidence that the differences cases evidence that the differences seem to go away around the same age seem to go away around the same age across kids (& across languages).across kids (& across languages).

It becomes interesting It becomes interesting to know…to know…

What are the principles that kids What are the principles that kids know as early as we can test?know as early as we can test?

What are the principles that are What are the principles that are delayed, and until when are they delayed, and until when are they delayed?delayed?

Wexler (1997) suggests the Wexler (1997) suggests the terminology terminology Continuous DevelopmentContinuous Development for this model (vs. for this model (vs. RigidityRigidity). (so, ). (so, *tadpole *tadpole frog) frog)

Is maturation a cop-Is maturation a cop-out?out?

If a kid doesn’t behave according to If a kid doesn’t behave according to Principle X of UG, we say that kid’s Principle X of UG, we say that kid’s grammar needs to mature until it gets grammar needs to mature until it gets Principle X. Can’t we just say that Principle X. Can’t we just say that about anything? about anything? Can we ever show that Can we ever show that “it just matures” is false?“it just matures” is false?

Actually, yes—if it matures, if it is Actually, yes—if it matures, if it is on a biological schedule, then on a biological schedule, then it it can’t really differ from language to can’t really differ from language to languagelanguage (at least to any greater (at least to any greater extent than, say, malnutrition can extent than, say, malnutrition can delay puberty).delay puberty).

HowHow different is a different is a kid’s grammar?kid’s grammar?

In principle, it In principle, it couldcould be quite be quite different. different. Tadpoles Tadpoles dodo become become frogs in the real, biological frogs in the real, biological worldworld..

But it But it seemsseems like what kids have like what kids have is pretty close to what adults is pretty close to what adults have, based on empirical studies—have, based on empirical studies—leading to the hypothesis that leading to the hypothesis that there is a close connection there is a close connection between kids grammars and adult between kids grammars and adult grammars…grammars…

UG-constrained UG-constrained maturationmaturation

Borer & Wexler (1992) introduced the Borer & Wexler (1992) introduced the hypothesis as hypothesis as UG-Constrained UG-Constrained MaturationMaturation, which says that , which says that all all child grammatical representations child grammatical representations are representations that are are representations that are available in UGavailable in UG..

In other words, a kid’s syntactic In other words, a kid’s syntactic tree is one that could exist in some tree is one that could exist in some adult language without violating adult language without violating principles of UG.principles of UG.

UG-constrained UG-constrained maturationmaturation

This hypothesis only allows for This hypothesis only allows for certain kinds of “kid deficits”—certain kinds of “kid deficits”—a kid grammar can rule out a a kid grammar can rule out a structurestructure, which an adult , which an adult (speaking (speaking somesome adult language) adult language) would consider grammatical, but would consider grammatical, but it cannot it cannot allowallow a structure that a structure that nono adult language would make adult language would make grammaticalgrammatical..

Optional infinitivesOptional infinitives

Young, Young, youngyoung kids show evidence of kids show evidence of knowing how to inflect, move verbs, knowing how to inflect, move verbs, etc. They know the parameter etc. They know the parameter settings for their language, even. settings for their language, even. Kids know a Kids know a lotlot..

ButBut—kid allows nonfinite forms in —kid allows nonfinite forms in contexts that adult requires finite contexts that adult requires finite forms in.forms in.

How does this fit in to UGCM?How does this fit in to UGCM?

Optional infinitivesOptional infinitives Take the Wexler (1998) view that kids Take the Wexler (1998) view that kids don’t know that D is don’t know that D is interpretableinterpretable. . This can be seen as another kind of This can be seen as another kind of “coordination” issue—coordinating the “coordination” issue—coordinating the syntactic system and the syntactic system and the interpretation system.interpretation system.

As long as the As long as the syntactic syntactic system system doesn’t require T or Agr, this fits doesn’t require T or Agr, this fits in with UGCM.in with UGCM.

That is, we take “Have T” and “Have That is, we take “Have T” and “Have Agr” as being principles Agr” as being principles outsideoutside the the syntax—maybe tied to discourse.syntax—maybe tied to discourse.

PassivesPassives John kicked the ballJohn kicked the ball (active) (active) The ball was kicked (by John)The ball was kicked (by John) (passive) (passive)

Standard analysis: Standard analysis: the ballthe ball starts off starts off as complement of V in both; in the as complement of V in both; in the passive, the agent is suppressed and passive, the agent is suppressed and the verb is deprived of its ability to the verb is deprived of its ability to assign Case. Thus, assign Case. Thus, the ballthe ball moves into moves into SpecIP to get Case.SpecIP to get Case.

The ballThe ballii was kicked t was kicked tii..

PassivesPassives

The ballThe ballii was kicked t was kicked tii..

The chain between The chain between the ballthe ball and and tt created by moving created by moving the ballthe ball into into SpecIP is an SpecIP is an A-chainA-chain (a chain whose (a chain whose top is in a position where you can top is in a position where you can only find arguments). Like subject only find arguments). Like subject position (SpecTP or SpecAgrSP).position (SpecTP or SpecAgrSP).

Kids vs. passives…Kids vs. passives…

It was observed early on (Horgan 1978, It was observed early on (Horgan 1978, Maratsos et al. 1985) that kids have Maratsos et al. 1985) that kids have trouble with passives.trouble with passives.

But there are a couple of asymmetries:But there are a couple of asymmetries: Kids are better at actional passives than Kids are better at actional passives than nonactional passives:nonactional passives:

Jasmine was combed (by Wendy)Jasmine was combed (by Wendy) Peter Pan was feared (by Captain Hook)Peter Pan was feared (by Captain Hook)

Kids are better at short passives (without Kids are better at short passives (without the the byby-phrase) earlier than long passives.-phrase) earlier than long passives.

Why are kids better at Why are kids better at actional passives?actional passives?

In English at least, it seems In English at least, it seems like there are two kinds of like there are two kinds of words with passive morphology:words with passive morphology:

VerbalVerbal:: The suspect was seen.The suspect was seen. AdjectivalAdjectival:: His hair seems His hair seems combed.combed.

Verbal and adjectival Verbal and adjectival passivespassives

Generally, non-action verbs make poor adjectival Generally, non-action verbs make poor adjectival passives (while action verbs are fine):passives (while action verbs are fine):

*The suspect seems seen. The seen suspect *The suspect seems seen. The seen suspect (fled). Seen though the movie was, John went to (fled). Seen though the movie was, John went to see it again.see it again.

The cloth seems torn. The torn cloth (is The cloth seems torn. The torn cloth (is useless). Torn though the cloth was, John used useless). Torn though the cloth was, John used it anyway.it anyway.

So: Action verbs can form adjectival passives.So: Action verbs can form adjectival passives. Conclusion: It should be possible for kids to Conclusion: It should be possible for kids to say/understand passive-like action verbs, but say/understand passive-like action verbs, but not passive-like non-action verbs.not passive-like non-action verbs.

Verbal vs. adjectival Verbal vs. adjectival passivespassives

Borer & Wexler (1987): the early passives Borer & Wexler (1987): the early passives that we see kids produce/comprehend are that we see kids produce/comprehend are adjectival passives.adjectival passives.

The crucial difference (on B&W’s The crucial difference (on B&W’s analysis) between verbal and adjectival analysis) between verbal and adjectival passives has to do with passives has to do with wherewhere the the modification of the argument structure modification of the argument structure happens.happens. adjectival passive: adjectival passive: in the lexiconin the lexicon((turns it into a real adjectiveturns it into a real adjective))

verbal passive: verbal passive: in the syntaxin the syntax So, kids can’t do the syntactic passive. So, kids can’t do the syntactic passive. Why?Why?

Verbal vs. adjectival Verbal vs. adjectival passivespassives

The bottom line is:The bottom line is: verbal passives move their argument into verbal passives move their argument into the usual external argument positionthe usual external argument position

adjectival passives just adjectival passives just startstart their their argument in the usual external argument argument in the usual external argument positionposition

The The movement movement of the internal argument of the internal argument to the external position is the to the external position is the problem.problem.

Borer & Wexler (1987) propose what we can Borer & Wexler (1987) propose what we can call the call the A-chain Deficit HypothesisA-chain Deficit Hypothesis::A-chains are unavailable to kidsA-chains are unavailable to kidswith a “Proto-UG”.with a “Proto-UG”.

Adjectival passives and Adjectival passives and verbal passivesverbal passives So, the “passives” that we see young kids So, the “passives” that we see young kids

produce are actually deceptions. They are not produce are actually deceptions. They are not reallyreally verbal passives—and if the ACDH is verbal passives—and if the ACDH is right, they couldn’t be—but are adjectival.right, they couldn’t be—but are adjectival.

Looking at Hebrew, where adjectival passives Looking at Hebrew, where adjectival passives and verbal passives are distinct, B&W observe and verbal passives are distinct, B&W observe that adjectival passives are (clearly) used that adjectival passives are (clearly) used early, and verbal passive only appear at early, and verbal passive only appear at “school age” (Berman & Sagi 1981).“school age” (Berman & Sagi 1981).

There is an additional complication in Hebrew that we There is an additional complication in Hebrew that we won’t get into here, which involves the availability of won’t get into here, which involves the availability of unmoved Themes in verbal passives. B&W87 argued that kids unmoved Themes in verbal passives. B&W87 argued that kids were also lacking the case assignment mechanism that were also lacking the case assignment mechanism that allows this; we might alternatively think of it as being allows this; we might alternatively think of it as being like the Russian Genative of Negation discussed later, like the Russian Genative of Negation discussed later, involving an (optional) “hidden” movement.involving an (optional) “hidden” movement.

PredictionsPredictions

A-chain Deficit HypothesisA-chain Deficit Hypothesis::A-chains are unavailable to kidsA-chains are unavailable to kidswith a “Proto-UG”.with a “Proto-UG”.

So, suppose that this is true. What are the So, suppose that this is true. What are the predictions?predictions? Of course, (real, verbal) passives will be Of course, (real, verbal) passives will be impossible.impossible.

But also anything else with an A-chain.But also anything else with an A-chain. Ouch. Well, since the VPISH, pretty much Ouch. Well, since the VPISH, pretty much everyevery sentence sentence has an A-chain, so that can’t be right. It has to be has an A-chain, so that can’t be right. It has to be something special about raising the Theme. One something special about raising the Theme. One suggestion (B&W92) is “non-canoncial suggestion (B&W92) is “non-canoncial -role assignment”, -role assignment”, though that’s not great either, since we need to add though that’s not great either, since we need to add some kind of theory of what canonical some kind of theory of what canonical -role assignment -role assignment is.is.

Other things with A-Other things with A-chainschains

The VPISH has given us a hint that The VPISH has given us a hint that perhaps “A-chain” is not exactly the perhaps “A-chain” is not exactly the right concept, but let’s focus on the right concept, but let’s focus on the kind of object-to-subject movement that kind of object-to-subject movement that we see in passives.we see in passives.

Other obvious candidate: Other obvious candidate: UnaccusativesUnaccusatives..

This opens up a bigger can of worms. Do This opens up a bigger can of worms. Do kids have problems with unaccusatives? kids have problems with unaccusatives? What is the nature of the problems?What is the nature of the problems?

UnaccusativesUnaccusatives There are two kinds of intransitive There are two kinds of intransitive verbs:verbs: UnergativeUnergative (subject-type argument)(subject-type argument) UnaccusativeUnaccusative (object-type argument)(object-type argument)

The The unergativeunergative verbs have an verbs have an external external argument— argument— just like a transitive verbjust like a transitive verb..

The The unaccusative unaccusative verbs have only an verbs have only an internal internal argument, which moves to argument, which moves to subject position—subject position—just like in a just like in a passivepassive..

Unaccusatives ≈ Unaccusatives ≈ passivespassives

An unaccusative is structurally like a An unaccusative is structurally like a passive:passive: The trainThe trainii arrived arrived ttii..

An unergative is not.An unergative is not. The baby giggled.The baby giggled.

So we expect kids to have the same So we expect kids to have the same troubles with unaccusatives and passives.troubles with unaccusatives and passives.

In particular, we expect kids to have In particular, we expect kids to have no no way to represent an unaccusativeway to represent an unaccusative..

But we But we knowknow kids use and understand verbs kids use and understand verbs that are, for adults, unaccusative. So that are, for adults, unaccusative. So what is the implication?what is the implication?

S-homophonyS-homophony Borer & Wexler suggest that what’s Borer & Wexler suggest that what’s happening when a kid comprehends/uses happening when a kid comprehends/uses an “unaccusative” verb is that the an “unaccusative” verb is that the verb is misanalyzed as an unergative.verb is misanalyzed as an unergative.

It has to be, the kid—by hypothesis—It has to be, the kid—by hypothesis—can’t represent an unaccusative can’t represent an unaccusative structure.structure. The boat sank.The boat sank. *The boat*The boatii sank sank ttii.. The doll giggled.The doll giggled.

The reason this happens is that the The reason this happens is that the surface form doesn’t distinguish surface form doesn’t distinguish between unergatives and unaccusatives. between unergatives and unaccusatives. They are “S(yntactic)-homophones.”They are “S(yntactic)-homophones.”

S-homophonyS-homophony That is: That is: the (immature) kid can’t tell the (immature) kid can’t tell the difference between an unergative the difference between an unergative and an unaccusativeand an unaccusative..

Is there evidence of that?Is there evidence of that? Maybe, there’s some. But there’s also some Maybe, there’s some. But there’s also some evidence against it.evidence against it.

Is this even conceptually a good idea?Is this even conceptually a good idea? Probably not. Why is an unaccusative Probably not. Why is an unaccusative unaccusative? Because the argument is a unaccusative? Because the argument is a Theme. UTAH says Themes are in object Theme. UTAH says Themes are in object position. So when a kid uses position. So when a kid uses sinksink or or fallfall do they think the argument is an Agent? Or do they think the argument is an Agent? Or do they violate UTAH? And once their do they violate UTAH? And once their grammar matures, how do they recover?grammar matures, how do they recover?

B&W: Pro-conflation: B&W: Pro-conflation: CausativesCausatives

Causativization adds a causative argument Causativization adds a causative argument (in English, it happens to be Ø):(in English, it happens to be Ø): Mom’s favorite vase broke.Mom’s favorite vase broke. Timmy broke Mom’s favorite vase.Timmy broke Mom’s favorite vase.

In English (not in all languages, e.g., In English (not in all languages, e.g., Hebrew), Hebrew), this can only happen if there this can only happen if there wasn’t already an external argumentwasn’t already an external argument. Works . Works for unaccusatives, but not for unergatives for unaccusatives, but not for unergatives or transitives:or transitives: The doll giggled.The doll giggled. *Peter giggled the doll.*Peter giggled the doll. Peter kicked the ball.Peter kicked the ball. *I kicked Peter the ball (‘I made Peter kick the *I kicked Peter the ball (‘I made Peter kick the ball.’)ball.’)

CausativesCausatives If kids can’t represent unaccusatives (that If kids can’t represent unaccusatives (that is, if all intransitives are for them is, if all intransitives are for them unergative), then they can’t make that unergative), then they can’t make that distinction. distinction.

Kids hear:Kids hear: The door opened. Daddy opened the doorThe door opened. Daddy opened the door..

The kids cannot reach the (adult-)correct The kids cannot reach the (adult-)correct conclusion that causativization only works conclusion that causativization only works for unaccusatives. It must be possible for for unaccusatives. It must be possible for any intransitive.any intransitive.

And indeed, kids over-apply causativization And indeed, kids over-apply causativization to unergatives too:to unergatives too: Daddy giggled the doll.Daddy giggled the doll.

Anti-conflation: Kim Anti-conflation: Kim (1997)(1997)

Kim (1997) observed that in Korean, Kim (1997) observed that in Korean, kids make a “negation misplacement” kids make a “negation misplacement” error error onlyonly with respect to objects with respect to objects and unaccusative subjects, never to and unaccusative subjects, never to unergative or transitive subjects:unergative or transitive subjects: na na anan pap mek-e pap mek-e ((adultadult: na pap(-ul) : na pap(-ul) an mek-e)an mek-e)I I negneg rice eat rice eat‘I do not eat rice.’‘I do not eat rice.’

anan ippal ssek-e ippal ssek-e ((adultadult: ippal(-i) an : ippal(-i) an ssek-e)ssek-e)negneg teeth rot teeth rot‘I won’t have a cavity.’‘I won’t have a cavity.’

Anti-conflation: Guasti Anti-conflation: Guasti (2002)(2002)

Guasti also notes (in the textbook, without Guasti also notes (in the textbook, without any citation of any other study) that Italian any citation of any other study) that Italian kids generally get the auxiliary selection kids generally get the auxiliary selection right—much earlier than the purported right—much earlier than the purported maturation.maturation. Gianni Gianni {è, *ha} andato.{è, *ha} andato. (adults)(adults)Gianni {is, has} leftGianni {is, has} left‘Gianni left.’‘Gianni left.’

Diana between 2;0 and 2;7 produced 22 relevant Diana between 2;0 and 2;7 produced 22 relevant sentences and 19 of them correctly had sentences and 19 of them correctly had bebe..

Guasti concludes that this is bad for the Guasti concludes that this is bad for the “maturational account”—but it’s really only “maturational account”—but it’s really only bad for the ACDH version of it. Something bad for the ACDH version of it. Something else else could still be maturing.could still be maturing.

Pro-conflation: Pro-conflation: Babyonyshev et al. Babyonyshev et al.

(1998)(1998) Testing the idea from Borer & Wexler Testing the idea from Borer & Wexler (1987) that unaccusatives are (1987) that unaccusatives are analyzed as if they are unergatives analyzed as if they are unergatives by kids in the pre-passive stage of by kids in the pre-passive stage of life.life.

Turns out that Russian provides a Turns out that Russian provides a nice test of nice test of unaccusativity/unergativity with the unaccusativity/unergativity with the “genitive of negation” so we can “genitive of negation” so we can directlydirectly check to see how kids are check to see how kids are analyzing their intransitives.analyzing their intransitives.

Russian genitive of Russian genitive of negationnegation In negative sentences, an object in the scope In negative sentences, an object in the scope

of negation can be of negation can be accusativeaccusative (if the object (if the object is definite/specific) is definite/specific) oror genitivegenitive (if the (if the object is indefinite/non-specific).object is indefinite/non-specific).

So: So: ability to be marked with genitive a ability to be marked with genitive a property of VP-internal indefinite objects.property of VP-internal indefinite objects. Ja ne poluchil pis’ma.Ja ne poluchil pis’ma.I not received letter-acc.plI not received letter-acc.pl‘I didn’t receive the/some letters.’‘I didn’t receive the/some letters.’

Ja ne poluchil (nikakix) pisem.Ja ne poluchil (nikakix) pisem.I not received (neg-kind-gen.pl) letter-gen.plI not received (neg-kind-gen.pl) letter-gen.pl‘I didn’t receive any letters.’‘I didn’t receive any letters.’

Ja poluchil pis’ma/*pisem.Ja poluchil pis’ma/*pisem.I received letter-acc.pl/*letter-gen.plI received letter-acc.pl/*letter-gen.pl‘I received the/some letters.’‘I received the/some letters.’

Russian genitive of Russian genitive of negationnegation

Arguments of Arguments of unaccusativesunaccusatives and and passivespassives (pronounced in their (pronounced in their postverbal, VP-internal base postverbal, VP-internal base position) can be marked with GoN.position) can be marked with GoN.

A small class of verbs A small class of verbs requiresrequires its arguments to be marked with its arguments to be marked with GoN (regardless of definiteness); GoN (regardless of definiteness); includes existential includes existential bebe..

Russian genitive of Russian genitive of negationnegation

Base-generated objects (arguments of Base-generated objects (arguments of passives and unaccusatives) still have passives and unaccusatives) still have a “hidden A-chain”, however. There is a “hidden A-chain”, however. There is some relation between these objects and some relation between these objects and the subject position that is (like?) an the subject position that is (like?) an A-chain.A-chain.

(They “move covertly”—it’s as if they move to (They “move covertly”—it’s as if they move to subject position, except that you pronounce the subject position, except that you pronounce the trace trace instead.)instead.)

We believe this based on the following We believe this based on the following facts about licensing of negative facts about licensing of negative phrases.phrases.

Covert movement of Covert movement of genitive argumentgenitive argument

Point 1:Point 1: When clausal negation When clausal negation co-occurs in the same clause co-occurs in the same clause with negative phrases, all is with negative phrases, all is well.well. [ [ anyany .. neg ], [ … neg … .. neg ], [ … neg … anyany]]

Point 2:Point 2: Negation in a lower Negation in a lower clause can’t license a negative clause can’t license a negative phrase in the upper clause.phrase in the upper clause. * [ * [ anyany … [ … neg … ]] … [ … neg … ]]

Covert movement of Covert movement of genitive argumentgenitive argument

Point 3:Point 3: A raised negative A raised negative phrase subject has to raise phrase subject has to raise toto a clause with negation—not a clause with negation—not fromfrom a clause with negation.a clause with negation. [ [ anyanyii … neg … [ … neg … [ ttii … ] ] … ] ]

* [ * [ anyanyii …[ …[ ttii … ] ] … ] ]

Covert movement of Covert movement of genitive argumentgenitive argument

Point 4:Point 4: A raising verb embedding a A raising verb embedding a clause with an unaccusative and an clause with an unaccusative and an genitive negative phrase needs to genitive negative phrase needs to have negation have negation above itabove it and and notnot down down with it.with it. [ … neg … [ … [ … neg … [ … anyany-gen … ]]-gen … ]] * [ … [ … neg … * [ … [ … neg … anyany-gen … ]]-gen … ]]

GoN acts as if it moved into the GoN acts as if it moved into the upper clause, we just can’t see it upper clause, we just can’t see it (it’s covert).(it’s covert).

Now, what do we expectNow, what do we expectpre-A-chain kids to do?pre-A-chain kids to do?

In GoN constructions, the In GoN constructions, the unaccusative argument is unaccusative argument is pronounced in its base-positionpronounced in its base-position There can be no re-analysis as an There can be no re-analysis as an unergative.unergative.

No S-homophones.No S-homophones. Moreover, GoN is prohibited with Moreover, GoN is prohibited with unergatives.unergatives.

This is pretty much impossible to This is pretty much impossible to solve—the kid’s stuck, and we solve—the kid’s stuck, and we expect them just not to use GoN.expect them just not to use GoN.

Testing the GoNTesting the GoN GoN is allowed with GoN is allowed with transitivestransitives and and these do not involve problematic A-these do not involve problematic A-chains.chains.

First order of business is to see if First order of business is to see if kids know how to use GoN in the kids know how to use GoN in the unproblematic cases.unproblematic cases.

Tested 30 kids in Moscow between 3;0 Tested 30 kids in Moscow between 3;0 and 6;6.and 6;6.

First result:First result: Kids use genitive about Kids use genitive about 75%75% of the time where it should be of the time where it should be used, around used, around 4%4% of the time where it of the time where it shouldn’t. Smart kids.shouldn’t. Smart kids.

Testing the GoNTesting the GoN Second result, split by age:Second result, split by age: Verbs that require Verbs that require GoN showed significant difference by age: GoN showed significant difference by age: younger kids (4;0) used GoN 30% of the time, younger kids (4;0) used GoN 30% of the time, older kids (5;4) used it 60% of the time.older kids (5;4) used it 60% of the time.

This is still fairly course—it turns out that This is still fairly course—it turns out that if we look at the individual subjects, we will if we look at the individual subjects, we will find find all and only all and only the patterns the hypothesis the patterns the hypothesis predicts with respect to where kids accept GoN.predicts with respect to where kids accept GoN. At least this is what Babyonyshev et al. assert—it’s At least this is what Babyonyshev et al. assert—it’s actually not really clear that this is the case (Hale actually not really clear that this is the case (Hale 2001).2001).

Subject by subject use Subject by subject use of GoNof GoN

Kids divided by their case response Kids divided by their case response forfor transitive non-specifictransitive non-specific (adult: gen) (adult: gen) transitive specifictransitive specific (adult: acc) (adult: acc) unaccusativeunaccusative (adult: gen) (adult: gen) bleached unaccusativebleached unaccusative (adult: gen) (adult: gen)

They fell into classes.They fell into classes. Kids who Kids who don’t know how to use GoNdon’t know how to use GoN at all. at all. Kids who Kids who use GoN like adultsuse GoN like adults (post-A-chain (post-A-chain kids)kids)

Kids use Kids use GoN right for transitives, not for GoN right for transitives, not for unaccusativesunaccusatives..

*Kids use *Kids use GoN right for unaccusatives not GoN right for unaccusatives not for transitivesfor transitives..

Kids who use GoN right Kids who use GoN right for transitives, not for transitives, not for unaccusativesfor unaccusatives 7 7 reallyreally act as predicted: act as predicted:

Nom for both bleached and non-bleached Nom for both bleached and non-bleached unaccusatives. (Adults would have gen here; and unaccusatives. (Adults would have gen here; and nom for unergatives, as these kids have)nom for unergatives, as these kids have)

3 get non-bleached unaccusatives (only) 3 get non-bleached unaccusatives (only) right:right: Gen for non-bleached, nom for bleached.Gen for non-bleached, nom for bleached.

Explanation: maybe these kids are in transition, or Explanation: maybe these kids are in transition, or maybe UTAH vs. ACDH are fighting, or maybe it’s just maybe UTAH vs. ACDH are fighting, or maybe it’s just performance errors.performance errors.

8 get bleached unaccusatives (only) right:8 get bleached unaccusatives (only) right: Nom for non-bleached, gen for bleached.Nom for non-bleached, gen for bleached.

Explanation: Explanation: bebe is in this class, overwhelming is in this class, overwhelming frequency, learned by rote? So, we ignore bleached.frequency, learned by rote? So, we ignore bleached.

GoN as a diagnosticGoN as a diagnostic So, it’s not So, it’s not reallyreally clear what we have clear what we have here. We have something like a tendency here. We have something like a tendency toward a problem with unaccusatives, for a toward a problem with unaccusatives, for a certain set of kids. The results were not certain set of kids. The results were not as clear-cut as one might have hoped for, as clear-cut as one might have hoped for, however.however. Perhaps this is a problem with GoN as a true Perhaps this is a problem with GoN as a true diagnostic of unaccusativity, particularly with diagnostic of unaccusativity, particularly with respect to the “bleached” verbs.respect to the “bleached” verbs.

Perhaps this is a problem with the premise Perhaps this is a problem with the premise itself: maybe pre-passive kids itself: maybe pre-passive kids don’tdon’t have the have the same problem with unaccusatives as with same problem with unaccusatives as with passives.passives.

In any event, the case for unaccusatives is less In any event, the case for unaccusatives is less clear.clear.

Two possible Two possible interpretationsinterpretations The ACDH says that the object-to-subject The ACDH says that the object-to-subject

movement required in a passive is movement required in a passive is problematic, and there is at least some problematic, and there is at least some evidence that points to problems with evidence that points to problems with unaccusatives too. But that movement is unaccusatives too. But that movement is not the only thing they have in common.not the only thing they have in common.

ACDH:ACDH: A-Chain Deficit Hypothesis A-Chain Deficit Hypothesis(no A-chains)(no A-chains)

EARH:EARH: External Argument Requirement External Argument Requirement Hypothesis (external arguments required)Hypothesis (external arguments required)

Passives and unaccusatives both fail Passives and unaccusatives both fail both. Transitives and unergatives both both. Transitives and unergatives both pass both.pass both.

Possible support for Possible support for EARH over ACDHEARH over ACDH

Snyder, Hyams, and Crisma (1994) found that Snyder, Hyams, and Crisma (1994) found that French kids get auxiliary selection right from French kids get auxiliary selection right from a young age—in particular with reflexive a young age—in particular with reflexive clitics.clitics. Although the unaccusative/unergative distinction Although the unaccusative/unergative distinction seems to play a role in the selection of the seems to play a role in the selection of the auxiliary, it’s not a 1-to-1 correlation auxiliary, it’s not a 1-to-1 correlation (particularly in French, it might be closer in (particularly in French, it might be closer in Italian, though). Only Italian, though). Only somesome unaccusatives take unaccusatives take bebe, , and a kid still needs to figure out which.and a kid still needs to figure out which.

ReflexivesReflexives OTOH are much more reliable. There OTOH are much more reliable. There are good arguments for supposing that their are good arguments for supposing that their structure involves object-to-subject movement:structure involves object-to-subject movement: Le Le chienchienjj s sii’est [ t’est [ tii mordu t mordu tjj ] ]‘the dog bit itself.’‘the dog bit itself.’

EARH…EARH… If this analysis is right, then we have If this analysis is right, then we have “object-to-subject” movement just like in “object-to-subject” movement just like in passives and unaccusatives, yet passives and unaccusatives, yet kids kids cancan do do this at a young agethis at a young age. What gives?. What gives? There is an A-chain just like in unaccusatives and There is an A-chain just like in unaccusatives and passives. So the problem would seem not to be about passives. So the problem would seem not to be about A-chains.A-chains.

The reflexive and unaccusative/passive differ in The reflexive and unaccusative/passive differ in that the reflexives that the reflexives still have their external still have their external --role intactrole intact..

Hence:Hence: maybe the “pre-A-chain” kids are really maybe the “pre-A-chain” kids are really “obligatory external argument” kids (EARH).“obligatory external argument” kids (EARH).

What else does EARH What else does EARH predict?predict?

So, if EARH is right, it predicts kids So, if EARH is right, it predicts kids will do poorly on anything without an will do poorly on anything without an external argument. So far, we have:external argument. So far, we have: Fine: transitives, unergativesFine: transitives, unergatives Not fine: unaccusatives, passivesNot fine: unaccusatives, passives

What What else else lacks an external argument? lacks an external argument? Well, raising verbs and weather verbs:Well, raising verbs and weather verbs: JohnnyJohnnyii seems seems [ t[ tii to be riding a horsie to be riding a horsie ]].. It seems It seems [[ that Johnny is riding a that Johnny is riding a horsie horsie ]]..

It rained.It rained. So how do kids do on those?So how do kids do on those?

It rainedIt rained Kirby & Becker (submitted) looked at Kirby & Becker (submitted) looked at occurrences of expletive occurrences of expletive itit (among other (among other things).things). Kids use the referential pronoun Kids use the referential pronoun itit first. first. Kids leave out expletive Kids leave out expletive itit for a while. for a while. Then, kids use expletive Then, kids use expletive itit..

But that’s not even the important thing. But that’s not even the important thing. Kids Kids useuse the weather verbs (way too the weather verbs (way too early), and they have no external early), and they have no external argument. argument. Adam by 2;6, Eve by 1;12, Nina by 2;2, Peter Adam by 2;6, Eve by 1;12, Nina by 2;2, Peter by 2;6.by 2;6.

EARH seems to us to EARH seems to us to have troublehave trouble

Wexler (2004) reports that kids have trouble Wexler (2004) reports that kids have trouble with raising verbs:with raising verbs: BertBertii seems to Ernie [ seems to Ernie [ttii to be wearing a hat]. to be wearing a hat].

As predicted—they should have trouble with As predicted—they should have trouble with seemsseems generally (by EARH), or with the raising generally (by EARH), or with the raising itself (by ACDH). Except kids do great on:itself (by ACDH). Except kids do great on: It seems to Ernie [that Bert is wearing a hat].It seems to Ernie [that Bert is wearing a hat].

So, it’s not So, it’s not seemsseems that they have trouble that they have trouble with. We’re back to (something like) the ACDH with. We’re back to (something like) the ACDH again, it’s the again, it’s the movementmovement that matters. that matters. Except that we still at some point have to confront Except that we still at some point have to confront the Romance reflexives/auxiliaries fact.the Romance reflexives/auxiliaries fact.

ACDH seems to have ACDH seems to have troubletrouble

Except Becker (forthcoming, Except Becker (forthcoming, LILI) ) found that kids do fabulously with:found that kids do fabulously with: The hayThe hayii seems seems ttii to be on the ground. to be on the ground.

The dogThe dogii seemed seemed ttii to be purple. to be purple.

So, it looks like the So, it looks like the experiencerexperiencer is is causing problems, but only in causing problems, but only in raising structures (not raising structures (not itit structures):structures): #Bert#Bertii seems to Ernie [ seems to Ernie [ttii to be wearing to be wearing a hat].a hat].

Wexler (2004)Wexler (2004) Wexler (2004) proposed a new version of the Wexler (2004) proposed a new version of the ACDH relying on the concept of “phases” in ACDH relying on the concept of “phases” in minimalist syntax.minimalist syntax.

The basic idea of a phase is that a tree is The basic idea of a phase is that a tree is built up from bottom to top in “chunks” and built up from bottom to top in “chunks” and once a chunk has been built, you can’t “see once a chunk has been built, you can’t “see into it” any further than the edge:into it” any further than the edge: [[CPCP … … [[vvPP specifier specifier vv [[VPVP … ] … ] … ]… ] … ] … ]

Consequence: Consequence: unless an embedded DP can get unless an embedded DP can get into Specinto SpecvvP, it will be “frozen” inside the P, it will be “frozen” inside the vvP.P.

Proposal:Proposal: Some Some vvs for adults are “defective” s for adults are “defective” (not phases), including unaccusatives and (not phases), including unaccusatives and passives. For kids, no passives. For kids, no vvs are “defective”.s are “defective”.

Wexler (2004)Wexler (2004) Universal Phase RequirementUniversal Phase RequirementFor the immature child, For the immature child, vv alwaysalways defines a defines a phase.phase.

Effect of this is that movement of objects into Effect of this is that movement of objects into subject position is impossible. The object subject position is impossible. The object doesn’t go into Specdoesn’t go into SpecvvP and for kids, that’s P and for kids, that’s obligatory.obligatory.

One prediction this makes is that if there is One prediction this makes is that if there is some other reasonsome other reason for kids to get the object for kids to get the object into Specinto SpecvvP, then unaccusatives should be P, then unaccusatives should be possible.possible. One such reason would be if the object were a One such reason would be if the object were a whwh--word. The idea is that the word. The idea is that the whwh-word first moves to -word first moves to SpecSpecvvP and then moves on.P and then moves on.

So, let’s check…So, let’s check…

Hirsch & Wexler (2004)Hirsch & Wexler (2004)

Hirsch & Wexler looked into this, and Hirsch & Wexler looked into this, and discovered that indeed:discovered that indeed: #Bert seems to Ernie to be wearing a hat.#Bert seems to Ernie to be wearing a hat. Who seems to Ernie to be wearing a hat?Who seems to Ernie to be wearing a hat?

(I think—I must confess, I don’t actually (I think—I must confess, I don’t actually have this paper, this is based on secondhand have this paper, this is based on secondhand information and guesswork, but it’s probably information and guesswork, but it’s probably what they found.)what they found.)

But this is beautiful and awesome (cf. Ken’s But this is beautiful and awesome (cf. Ken’s talk at BUCLD 2004), what a weird thing to be talk at BUCLD 2004), what a weird thing to be true, but yet true, but yet predictedpredicted..

Hyams & Snyder (2005)Hyams & Snyder (2005) Great, but neither the ACDH nor the UPR Great, but neither the ACDH nor the UPR predicts the thing with the reflexive predicts the thing with the reflexive clitics in Romance. Kids should fail clitics in Romance. Kids should fail (because the object moves out), but they (because the object moves out), but they don’t. This was why EARH beat ACDH in the don’t. This was why EARH beat ACDH in the first place.first place.

Interesting, quite different idea, based on Interesting, quite different idea, based on a different analysis of the passive. a different analysis of the passive. “Smuggling” (Collins 2005).“Smuggling” (Collins 2005).

Basic idea: You can’t move A-move a DP over another Basic idea: You can’t move A-move a DP over another one, the solution is to first move something containing one, the solution is to first move something containing the DP over the other one, then move the DP out. Maybe the DP over the other one, then move the DP out. Maybe I’ll draw it on the board.I’ll draw it on the board.

I think this is the same concept as what Sauerland I think this is the same concept as what Sauerland (1994?) called “Surfing”(1994?) called “Surfing”

Bottom line: Children have trouble A-moving Bottom line: Children have trouble A-moving a DP past another argument.a DP past another argument.

Hyams & Snyder (2005)Hyams & Snyder (2005) Romance reflexives:Romance reflexives:

Suppose that Suppose that vv is the REFL morpheme, nothing is the REFL morpheme, nothing in Specin SpecvvP. Nothing to get in the way.P. Nothing to get in the way.

Passives:Passives: PRO or Agent of PRO or Agent of byby-phrase is in Spec-phrase is in SpecvvP. P. Smuggling required.Smuggling required.

Unaccusatives:Unaccusatives: ? Well, but are we ? Well, but are we suresure kids have trouble with kids have trouble with unaccusatives?unaccusatives?

Raising past experiencers:Raising past experiencers: Experiencer gets in the way. Smuggling Experiencer gets in the way. Smuggling required.required.

Raising without experiencers, raising Raising without experiencers, raising with with whwh-words over experiencers:-words over experiencers: No smuggling required.No smuggling required.

Beauty & awe: Lg. acq. as Beauty & awe: Lg. acq. as high sciencehigh science Notice what a nice progression we’ve had Notice what a nice progression we’ve had

here, even if we haven’t necessarily come to here, even if we haven’t necessarily come to a definitive conclusion.a definitive conclusion. Passives were hard? Why? The movement. What else Passives were hard? Why? The movement. What else has the movement?has the movement?

Unaccusatives? Maybe also hard. But Romance Unaccusatives? Maybe also hard. But Romance reflexives aren’t. So, maybe it’s the external reflexives aren’t. So, maybe it’s the external argument. What else lacks an external argument?argument. What else lacks an external argument?

Raising verbs? Hard, but only when there’s the Raising verbs? Hard, but only when there’s the movement. So, it’s the movement still, but it’s movement. So, it’s the movement still, but it’s not A-chains exactly, it’s the ability to get out not A-chains exactly, it’s the ability to get out of a phase. When might you independently be able of a phase. When might you independently be able to get out of a phase?to get out of a phase?

WhWh-words in raising contexts? Possible. But -words in raising contexts? Possible. But raising without experiencers seems to be raising without experiencers seems to be possible. Why?possible. Why?

Kids can’t smuggle? But we need to re-evaluate Kids can’t smuggle? But we need to re-evaluate unaccusatives.unaccusatives.

But are passives But are passives actually impossible?actually impossible? There’s some dispute about this, it turns There’s some dispute about this, it turns

out…out… It’s only getting the It’s only getting the -role to the -role to the by-by-phrase: phrase: Fox & Grodzinsky (1998)Fox & Grodzinsky (1998)

No, it’s not; kids think No, it’s not; kids think byby is is aboutabout: Hirsch & Wexler : Hirsch & Wexler (2005)(2005)

The tests were pragmatically ill-conceived: The tests were pragmatically ill-conceived: O’Brien, Grolla, Lillo-Martin (2005)O’Brien, Grolla, Lillo-Martin (2005)

There are languages with early passives. It There are languages with early passives. It depends on the properties of the child’s input.depends on the properties of the child’s input.

Inuktitut: Crago & Allen (1996)Inuktitut: Crago & Allen (1996) Sesotho: Demuth (1989?)Sesotho: Demuth (1989?)

Leads to a separate thread of argumentation: Those Leads to a separate thread of argumentation: Those things that look like “passives” in these languages things that look like “passives” in these languages aren’t really, cf. adjectival passives. See, e.g., aren’t really, cf. adjectival passives. See, e.g., footnote in Babyonyshev et al. (1998), Crawford footnote in Babyonyshev et al. (1998), Crawford (2003?)(2003?)

Fox, GrodzinskyFox, Grodzinsky Testing kids on actional/nonactional, Testing kids on actional/nonactional, long/short long/short bebe//get get passives:passives: Actional passives pose no problem for Actional passives pose no problem for comprehension (long or short).comprehension (long or short).

This is surprising—explanation, kids use This is surprising—explanation, kids use byby for for Affector, which turns out to be right, but it’s Affector, which turns out to be right, but it’s not due to transmitting the not due to transmitting the -role.-role.

Get Get passives (long) seem to pose no passives (long) seem to pose no problem.problem.

Nonactional short passives are pretty well Nonactional short passives are pretty well comprehended.comprehended.

Nonactional long passives are at chance.Nonactional long passives are at chance. Kids can’t transmit the Kids can’t transmit the -role to the -role to the byby-phrase.-phrase.

-transmission-transmission In verbal be passives, the In verbal be passives, the -role seems to -role seems to be “transmitted” to be “transmitted” to byby:: Aladdin is pushed by Jasmine (agent).Aladdin is pushed by Jasmine (agent). Captain Hook is feared by Michael (experiencer)Captain Hook is feared by Michael (experiencer) A cake is offered to Ariel by Pinocchio (source)A cake is offered to Ariel by Pinocchio (source) The ship was sunk [PRO to collect the insurance].The ship was sunk [PRO to collect the insurance].

But not with But not with getget-passives (-passives (byby works alone). works alone). The ship got sunk [PRO to collect the insurance].The ship got sunk [PRO to collect the insurance].

F&G suggest problem with F&G suggest problem with -transmission due -transmission due to processing (only option left is direct to processing (only option left is direct assignment from assignment from byby); for nonactional verbs, ); for nonactional verbs, get get passive).passive).

Hirsch & Wexler (2005)Hirsch & Wexler (2005)

Tested kids on Tested kids on byby phrases vs. phrases vs. aboutabout phrases, as in:phrases, as in: The story about Elmo had cars in it.The story about Elmo had cars in it. The story by Elmo had cars in it.The story by Elmo had cars in it.

Kids treated Kids treated byby phrases like phrases like about about phrases. They did great on the phrases. They did great on the aboutabout ones (91%) and lousy on the ones (91%) and lousy on the byby ones ones (28%).(28%).

Conclusion: The F&G story about Conclusion: The F&G story about Affectors doesn’t seem to be right.Affectors doesn’t seem to be right.

O’Brien et al. (2005)O’Brien et al. (2005) Experiments showing that passive wasn’t good Experiments showing that passive wasn’t good were pragmatically flawed. Fix the flaw, fix were pragmatically flawed. Fix the flaw, fix the performance.the performance. Bill was kicked by Pete.Bill was kicked by Pete. (actional, fine)(actional, fine) Bill was seen by Pete.Bill was seen by Pete. (non-actional, (non-actional, problems)problems)

Except that lots of people can see, even if Except that lots of people can see, even if only one person is kicking. Plus, this is kind only one person is kicking. Plus, this is kind of weird if there isn’t a character to contrast of weird if there isn’t a character to contrast with Pete.with Pete.

Tried setting up the scenario better so that Tried setting up the scenario better so that the question of whether Bill was seen by Pete the question of whether Bill was seen by Pete was at issue—and 4-year olds got it 82% of the was at issue—and 4-year olds got it 82% of the time.time.

Demuth (1989), Allen & Demuth (1989), Allen & Crago (1996)Crago (1996)

Passives seem early in Sesotho & Inuktitut. Passives seem early in Sesotho & Inuktitut. The claim is that they’re frequent in the The claim is that they’re frequent in the input, they’re verbal (they involve an A-input, they’re verbal (they involve an A-chain), and they’re in use by age 3.chain), and they’re in use by age 3.

This is a problem for the Maturation This is a problem for the Maturation hypothesis as an explanation for the delay hypothesis as an explanation for the delay in passives. Age of maturation can’t vary in passives. Age of maturation can’t vary by language.by language. Two possible counters to this: 1) Those weren’t Two possible counters to this: 1) Those weren’t really A-chain-containing passives (Crawford really A-chain-containing passives (Crawford 2004); 2) It wasn’t ACDH after all, and Sesotho 2004); 2) It wasn’t ACDH after all, and Sesotho differs in the relevant respect from English. differs in the relevant respect from English. For example: smuggling not required? (For For example: smuggling not required? (For another day perhaps)another day perhaps)