GRS LX 700 English wh-questions Language Acquisition and What will John bake? Linguistic Theory Two components to forming a (main clause) wh-question (in English): Week 12. Move a wh-word to SpecCP. Acquirers and questions Move T to C (Subject-Aux Inversion—SAI)
Question formation Wh-inversion→Wh-fronting
Declarative: John will buy coffee. English, German: Both. What will John buy? Wh-inversion: What will John buy? Japanese Korean: neither. Wh-fronting: What will John buy? John will buy what? Yes/No-inversion: Will John buy coffee? Finnish: Wh-fronting only. What John will buy? Greenberg (1963): Unattested: Wh-inversion only. Wh-inversion implies Wh-fronting. *Will John buy what? Yes/No-inversion implies Wh-inversion.
Y/N-inversion→Wh-inversion Universals and parameters
English: Both Even if it’s not completely clear what accounts Will John buy coffee? What will John buy? for the implicational universals, inversion and wh-fronting do seem to be independent. Japanese: Neither John will buy coffee? John will buy what? A kid needs to learn what his/her language does in each domain. Lithuanian: Wh-inversion only. Wh-inversion implies Wh-fronting: Perhaps the only John will buy coffee? What will John buy? reason you’d move T to C is to get a [wh] feature Unattested: Y/N-inversion only. originally on T into a position where it can be checked by a wh-word in SpecCP (Wh-criterion, see Guasti). Will John buy coffee? What John will buy? Y/N-inversion implies Wh-inversion: ?
1 Kids get these parameters Kuczaj & Maratsos (1983) down early
Guasti (2000): Adam, Eve, and Sarah pretty Form Abe Ben much never left wh-words in situ, and when Kids seem to they did it was generally in a (grammatical) echo Uninv Inv Uninv Inv learn question. can 2;5 2;11 2;6 2;10 auxiliaries Same with inversion, there seem to be very few is (cop) 2;7 3;1 2;4 2;8 one by one; (on the order of 1%) errors of non-inversion in are (cop) 2;9 3;0 2;7 2;10 they appear German, Italian, Swedish. is (aux) 3;0 3;0 2;7 3;1 at different Yet Bellugi (1971)—very famously—seemed to find something different in English… Stages: are (aux) 3;0 3;1 2;10 3;0 times. SAI in yes-no questions, not in wh-questions will 3;0 3;1 2;10 2;10 Notice this runs counter to Greenberg’s univeral. SAI in positive questions, not in negative questions.
Kuczaj & Maratsos (1983) Kuczaj & Maratsos (1983)
Form Abe Ben Each Form Abe Ben Only correctly Uninv Inv Uninv Inv auxiliary Uninv Inv Uninv Inv inverted can 2;5 2;11 2;6 2;10 seems be first can 2;5 2;11 2;6 2;10 verbs is (cop) 2;7 3;1 2;4 2;8 used outside is (cop) 2;7 3;1 2;4 2;8 (auxiliaries) are (cop) 2;9 3;0 2;7 2;10 of inversion are (cop) 2;9 3;0 2;7 2;10 appear in is (aux) 3;0 3;0 2;7 3;1 contexts, is (aux) 3;0 3;0 2;7 3;1 child speech are (aux) 3;0 3;1 2;10 3;0 only later in are (aux) 3;0 3;1 2;10 3;0 (no inversion will 3;0 3;1 2;10 2;10 inversions will 3;0 3;1 2;10 2;10 of main verbs)
A famous non-result: SAI in A famous non-result: SAI in YNQs before SAI in whQs YNQs before SAI in whQs Problem is, seems to be true of Adam’s Adam: At a certain files, but not true generally… YNQs WhQs point, inversion Inv Uninv Inv Uninv appears in yes-no Several later studies with better sampling 3;0 0 1 0 3 questions—but show no identifiable stage where yes-no 3;5 198 7 9 22 inversion with wh- questions invert while wh-questions 3;8 33 5 questions is still don’t—in fact, even the frequency doesn’t infrequent. Soon go in one direction for all kids. afterwards, inversion is frequent for both types of questions.
2 Stromswold (1990, table 5.5) Conclusion really seems to be % of inversion WHQ vs.YNQ Kids will sometimes fail to invert. Child WH YN Child WH YN Kids will sometimes fail to invert more in Adam 88.3 96.6 Nathan 60.1 46.2 one construction (e.g., wh-questions) than Allison 85.7 100 Nina 98.5 93.9 in another (e.g., yes/no-questions), but April 91.7 94.1 Peter 92.1 98.5 which one gets the advantage seems to Eve 95.5 87.2 Ross 99.3 97 vary by kid. Mark 97.9 97.6 Sarah 92.9 91.9 Naomi 96.2 94.2 Shem 95.6 79 MEAN 93 93.7
SAI errors: doubling Doubling errors
A double-auxiliary error, both an inverted Are the kids pronouncing a “loud trace” and an un-inverted auxiliary: of (head-)movement? (Are they moving Why did you did scare me? the auxiliary but failing to leave the trace How can he can look? unpronounced?) That would be interesting. A “double-tensing” error (where an Are they just forgetting what they are auxiliary moves to I but the verb surfaces trying to say midway through and with tense). “blending” two structures? (one with and What did you bought? one without movement) What did you did?
Nakayama (1987) Inversion in negation
Guasti, Thornton & Wexler (BUCLD 1995) The longer the subject is, the more likely a looked at doubling in negative questions. kid is to make a doubling error; the length of the VP makes no difference. Previous results (Bellugi 1967, 1971, Is [the boy who is watching Mickey] is happy? Stromswold 1990) indicated that kids tend Looks like blending, rather than the (more to invert less often in negative questions. interesting) “loud trace” idea… Common First: True? error type: Second: Why? Is [the boy who is watching M], is he happy?
3 GTW (1995) GTW (1995)
Elicited negative questions… Kids got positive questions right for the I heard the snail doesn’t like some things to most part. eat. Ask him what. 88% of kids’ wh-questions had inversion There was one place Gummi Bear couldn’t eat 96% of kids’ yes-no questions had inversion the raisin. Ask the snail where. Except youngest kid (3;8), who had inversion only 42% of the time. One of these guys doesn’t like cheese. Ask the snail who. Kids got negative declaratives right without exception, with do-support and I heard that the snail doesn’t like potato chips. Could you ask him if he doesn’t? clitic n’t.
GTW (1995) GTW (1995)
Kids got lots of negative wh-questions wrong. But kids got negative subject wh-questions right. Aux-doubling which one doesn’t like his hair messed up? (4;0) What kind of bread do you don’t like? (3;10) …as well as how-come questions. How come the dentist can’t brush all the teeth? (4;2) Neg & Aux doubling Why can’t she can’t go underneath? (4;0) Re: Not structure No I to C raising (inversion) Why can you not eat chocolate? (4;1) Where he couldn’t eat the raisins? (4;0) Kids only do this with object and adjunct wh- questions—if kids just sometimes prefer not instead of Not structure n’t, we would expect them to use it just as often with Why can you not eat chocolate? (4;1) subject wh-questions.
GTW (1995) GTW (1995)
So, in sum: The kids’ errors all seem to have the character of Kids get positive questions right keeping negation inside the IP. What did he didn’t wanna bring to school? (4;1) Kids get negative declaratives right What she doesn’t want for her witch’s brew? (3;8) Kids get negative subject questions right. Why can you not eat chocolate? (4;1) Kids get negative how-come questions right. Why can’t she can’t go underneath? (4;3) Kids make errors in negative wh- questions where inversion is required. GTW propose that this is a legitimate option; Where inversion isn’t required (or where citing Paduan (Italian dialect) as a language the sentence isn’t negative), they’re fine. doesn’t allow neg->C.
4 GTW (1995) “Auxless questions”
Re: subject and how come questions… Guasti (2002) discusses questions like In a subject question, we don’t know that the Where Daddy go? (Adam 2;3) subject wh-word got out of IP—maybe kids left it What I doing? (Eve 2;0) in IP… heck, maybe even adults do. By making some assumptions (inherited from Who left? Rizzi), Guasti finds these problematic. Wh- *Who did leave? movement requires SAI, so what moved to C? How come questions don’t require SAI in the Specifically, wh-movement depends on SAI, which adult language{./?} happens because [+wh] starts on T and must move to C so it can be in a Spec-head relation with the wh- How come John left? word in SpecCP. Also: subject questions need no *How come did John leave? inversion on this story.
Auxless questions Early, early wh-questions
Auxless questions are relatively common among There may be an early “formulaic” stage where wh-questions in the 2-4 age range. kids ask questions by just asking “Wh(’s) NP?”. Guasti/Rizzi’s suggestion: An auxiliary at the head of O’Grady (1997): “Because of their formulaic the root can be null (similar to the null subject story). character, it seems reasonable to treat these For adults, the head of the root is ForceP, but for kids it might be lower (FocP, where wh-words go). utterances as instantiations of a simple template rather than the product of whatever mechanism Kids who might otherwise say What I doing? will forms wh-questions in the adult grammar.” nevertheless not say Who laughing?. Subject wh- questions seem immune from “auxiliary drop.” But why? We already have lots of reason to think young kids know a lot about adult grammar by The Guasti/Rizzi explanation is pretty contrived, actually. The aux need not proceed as high as FocP then… What is simpler about a “simple for subject questions, so it ends up not being highest. template”? Not really any clear alternative, though…
Wh-subjects and wh-objects Early, early, early wh-questions
Is there a difference in the timing of Seidl and Hollich (2003) looked at emergence between subject wh-questions headturn preferences in really young kids. and object wh-questions? In English, there Minimizes demands of task is an apparent difference in complexity Use looking preferences to “answer” wh- (“distance” of movement, SAI). questions. What hit the apple? What did the apple hit? Where is the apple?
5 Seidl et al. Seidl et al.
Kids saw a little simplistic computer- Graph shows generated movie where, e.g., a book hit differences (target some keys. minus non-target). 20-month-olds Then there were two screens presented seemed quite capable side by side, one with a book displayed, of comprehending all one with keys displayed. three kinds. What hit the keys? (book) 15-month-olds What did the book hit? (keys) couldn’t do objects; Where is the book? (book) 13-month-olds couldn’t do any.
Processing, structural distance Processing, structural distance
The distance between the base and Re: preference for subject wh-questions; derived positions for an object wh-word is perhaps kids are sensitive to the number greater than the distance between the base of phrases a moving wh-phrase has to and derived positions for a subject wh- escape. This also makes other predictions: word. Whati will [IP Sue [VP read ti ]]?
Whati will [IP Sue [VP talk [PP about ti ]]]? Whati did [IP John [VP buy ti ]] ? Whati will [IP Sue [VP read [NP a book [PP about ti ]]]]? Whoi [IP ti [VP bought coffee ]] ?
Hildebrand (1987) But wait…
Tested (fairly old) kids on a paradigm of So kids make more errors extracting from more wh-questions of varying “depth” to see if deeply embedded structures. Is this a fact about more embedded wh-words are harder. the acquisition of wh-movement? Or is it just a fact about language processing in general? In a repetition task (4-10 year olds), it was almost uniformly true that the more What do adults do? deeply embedded the wh-word was, the My guess: Even for adults, the more complex more errors the kids made trying to repeat structures are (marginally) harder to process. it. Certainly true for subject vs. object relative clauses (the man who _ left vs. the man who I met _). Cf. NPAH later.
6 Does child wh-movement obey the Do kids have wh-traces in their adult rules for wh-movement? wh-questions? How do they perform on wanna- When the kids ask wh-questions, contraction? what structures are they using? Are Who do you want to help t? they like the adult structures? If not, Who do you wanna help t? how are they different? Are they Who do you want t to help you ? performing movement? Are there *Who do you wanna / t help you ? traces? Do the movements obey constraints (e.g., wh-island, ECP, …)? Crain & Thornton (1991) studied this…
Crain & Thornton (1991) Crain & Thornton (1991)
There are three guys in this story: Cookie The kids (2;10 to 5;5) all knew the wanna Monster, a dog, and this baby. One of contraction rule… them gets to take a walk, one gets to take a nap, and one gets to eat a cookie. The rat gets to choose who does each thing. So 59% of the time kids contracted to wanna one gets to take a walk, right? Ask Ratty with object questions (as allowed) who he wants. 4% of the time kids contracted to wanna Kid: Who do you want to take a walk? with subject questions (out for adult)
The ECP and argument-adjunct De Villiers, Roeper, and asymmetries Vainikka (1990) Moving a wh-word out of a wh-island is [Kid takes a shortcut home, rips dress, that better or worse depending on whether the night, kid tells parent about dress] wh-word is an argument (subject or object) When did she say t [she ripped her dress t]? or an adjunct. “at night” “that afternoon” When did she say t [wh how she ripped her dress t t ]? “at night” *“that afternoon”
*How did he ask [wh where to fix the car t ]? 3-6 year-olds allow short and long distance
What did he ask [wh how to fix t ] ? questions for complement clauses, don’t like long distance adjunct questions out of wh- islands…
7 De Villiers, Roeper, and Again, kids have a lot right—but Vainikka (1990) what do they have wrong? And kids make the argument-adjunct When kids make a mistake with a distinction the ECP makes for adults: question like… No wh-island, arguments/adjuncts both take long distance interpretation about 30- When did she say how she ripped her dress? 40% the time Argument wh-island, neither argument …it will often be that they answer nor adjuncts can move out (2-8% LD) something like “climbing over the fence”—answering the question How did Adjunct wh-islands, arguments can move she say t she ripped her dress? instead. out (30% LD) but not adjuncts (6% LD).
What are kids doing when they German partial wh-movement? answer a medial wh-word? Are they answering the last wh-word they Kids have been observed to produce questions saw? with an initial wh-word and a lower copy. Kids don’t answer medial wh-words in yes-no questions. What do you think what’s in her hat? Did Mickey tell Minnie what he bought? ‘What do you think is in her hat?’ What do you think where the marble is? Kids don’t answer wh-words in relatives. ‘Where do you think the marble is?’ How did you meet the man who sang? What do you think what Cookie Monster eats? ‘What do you think Cookie Monster eats?’
German partial wh-movement? Processing constraints?
Was hat er gesagt O’Grady (last year’s textbook) suggests that [ wie er das Kuchen machen kann ]? another reason why kids might answer the What has he said how he the cake make can intermediate wh-word is that they’ve already ‘How did he say he could make the cake?’ forgotten the matrix clause (citing Phinney 1981, Are kids treating the upper wh-word like a scope who found that 3-year olds often delete the marker? (Are they “speaking German”?) matrix subject and verb when repeating Hard to say with confidence, but it’s an interesting biclausal sentences). possibility. German partial wh-movement does have certain restrictions. Thornton (1990) and van Kempen (1997) showed that kids do this only out of finite Kids don’t answer a medial wh-word in a yes-no clauses, and German only allows partial movement out of finite clauses too. question, though..?
8 Speaking Irish? French? Speaking Irish? French?
Another crosslinguistic analogy we could make So, perhaps the kids’ non-adult use of is to Irish, French, and other languages that intermediate wh-words is actually a mis-analysis seem to show a certain amount of “wh- of English. agreement” when a wh-word passes through First, they suppose it is Irish, and the intermediate SpecCP. wh-words are the pronunciations of agreeing Ceapann tú go bhuailfidh an píobare an t-amhrán. complementizers. think you that play.fut the piper the song A medial wh-word is never a whole wh-phrase. A head? ‘You think that the piper will play the song.’ Then, they suppose it is French, and limit the Caidé aL cheapann tú aL bhuailfidh an píobare? agreement to subject wh-words. what WH think you WH play.fut the piper Sometimes production goes from S&O medial wh-questions ‘What do you think the piper will play?’ to just S. Then, they get to English. Je crois que Marie est partie. Qui crois-tu qui et partie?
Other constraints on wh- Superiority 3-5 movement from 3-5 year olds Adults: They reject adjunct extraction from NP Whoi ti slept where? *Howi did the mother see [his riding ti]? *Wherei did who sleep ti ? But they allow argument extraction…?
Whoi did the mother show [his copying ti] ? This is de Villiers’ example; seems ambiguous to me And the kids seem to have that down cold. between extraction and non-extraction readings. Better (Kid: It’s better if I start.) might be What did the mother show his eating? They reject adjunct extraction from rel. clause (from deVilliers and Plunkett, unpublished as of *How did [the woman who knitted t ] swim? i i 1995?) And reject extraction from temporal adjuncts
*Who did the elephant ask [before helping ti ]?
That-trace? That-trace?
Who did the pig believe that swam in the pond? Kids opt for the interpretation where the questions Some conflicting results? asks which, of the animals the pig believes, swam. Thornton (1990), production experiment Kids don’t go at all for the interpretation which entails a violation of that-trace (the pig believed that found that-trace violations 18% of the time who swam) subject wh-questions were used. (Phinney 1981)
McDaniel, Chiu and Maxfield (1995) found This is sort of mysterious, since languages differ an acceptance rate of 24% for that-trace as to whether they respect the that-trace filter. effects.
9 Grammar vs. Preferences Questioning out of quotations
These experiments are really testing preferences Adult languages generally can not not grammaticality. If they prefer the that-less question out of a quotation: variant, we won’t see that-trace violations even if they are strictly grammatical for the kid. *Whati did the boy say “Can I bring ti” ? Just because a structure is dispreferred (for whatever reason—frequency, difficulty, etc.) does But English, French and German kids (3-6 not mean that it is ungrammatical in the child’s years) seem to allow it. grammar. Preferences are not the best route to discovering Why? the properties of child grammar, though it’s hard to design grammaticality judgment experiments..
Correlates to questioning out False beliefs of quotations Kids before a certain age (usually before 4) seem Kids may not quite grasp the quotation yet. unable to take another person’s perspective: A significant proportion of kids around the same age range allow co-reference between a Little rabbit puts carrot in red basket, leaves. pronoun in the quotation and the subject: Mother rabbit comes in, moves carrot to blue basket. Little rabbit comes back. Where does he look for the carrot? “Hei can sit here” said Mickeyi.
Some kids will answer “the blue Perhaps, it has more to do with the fact that it basket”—unable to see that the little rabbit requires “getting into someone else’s head”… shouldn’t have known.
False beliefs & quotations False beliefs & quotations
Those same kids who answered “blue basket” So, perhaps it is understanding what a were also those who would do this: quotation is that is allowing kids to extract from them—they treat a quotation as a Mother bought cake, but wanted to surprise girl. regular clausal complement. When asked, mother claimed to have bought paper towels. What did Mother say she bought?
The “blue basket” kids answer “cake.”
10 Weak islands Weak islands
In the adult language, there is a certain In weak islands the implication fails: configuration which seems to create an Negation: island for movement of wh-adjuncts, John didn’t say Mary was coming by train. which arguably has to do with the logical John didn’t say Mary was coming. meaning. Factives: John forgot Mary was coming by train. Coming by train is a subset of the events coming. John forgot Mary was coming. With quantificational adverbs: John said Mary was coming by train implies John said Mary was coming. John often eats grapes with a fork. John often eats grapes.
Weak islands Weak islands
And in those cases, you can’t extract wh-adjuncts Four-year-olds have been observed to fail on the in the adult language. implication: Jim forgot that his aunt was arriving by train, so he went to the bus station to pick her up… Did Jim Whyi did John say (ti) that Mary left (ti)? forget that his aunt was coming? —Yes! Whyi did John forget (ti) that Mary left (*ti)?
Guess: They haven’t gotten the implication Whyi didn’t John say (ti) that Mary left (*ti)? pattern down for these non-monotonic- increasing environments. Whyi does John often say (ti) that Mary left (*ti)?
Weak islands Philip and de Villiers (1992)
Now: If kids haven’t gotten the Kids never allow LD association out of a wh- implication pattern, and if the implication island (they obeyed the purely syntactic pattern is implicated in the islandhood, do constraint). kids fail to observe weak islands just when *Whyi did the mother ask [what he made ti ]? they also fail on the implication pattern? The other facts were “generally in support”(de Villiers 1995) of the conclusion that where kids fail to make the inferences required by non- Philip and de Villiers (1992) looked into monotone-increasing environments, they also this… fail to treat them as movement islands.
11 Multiple questions Grebenyova (2005)
A fair amount of theoretical work has Russian as a multiple-movement language: concerned the treatment of multiple wh- chto kuda Smurf polozhil? questions. What where S put? Interpretation: E.g., the wh-typology: English (move one) vs. Japanese (move none) vs. Bulgarian (move PL (Pair-list): Who invited who for dinner? all). SP (Single pair): Which diplomat invited which journalist? Who invited the roommate of who for What do kids do with them? dinner? Well, but that’s lunacy—adults barely use Who invited who for dinner? them, how are we going to find out about English, Russian: PL, *SP kids? Serbo-Croatian, Japanese: PL, SP:
Grebenyova (2005) Grebenyova (2005) Attempts to elicit multiple interrogatives. Ok, let’s check CHILDES (parental speech). Story: 3 characters each hide a different thing. Varvara (1;7-2;11). Characters and items not in a natural category 737 single questions. Avoiding: Which x hid which Y? Who hid which X? Which x hid what? 1 multiple question. Add a character who doesn’t hide anything (and kto tebe chto podaril ? pointing that out). Who you what gave? nom acc Avoiding: What did everyone hide? Not mentioning the names of the characters in the lead-in Not very much input here. Avoiding: What did they hide? First time: single question. Decide to ask a more difficult question next time.
Grebenyova (2005) Grebenyova (2005) Tried non-subjects and adjuncts to figure out more And it worked: Kids (and adult controls) about the syntax: produced multiple wh-questions in PL contexts Who hid what? (but not SP contexts) about a third of the time in Who did Lizard give what? English, about half the time in Russian. Who did the dog find where? Syntax: English kids did it like adults. Russian Found some wh-in-situ for kids, both notably both for kids and adults found about two-thirds multiple kids 15% of the time did it like English fronting and one-third partial fronting: kids/adults: Kogo sobaka gde nashia? *Kto sprjatal chto? Who dog where found Who hid what Perhaps (for wh-in-situ; but partial fronting?) Acquisition of focus? Mixed/confusing input (which phrases can stay in situ)?
12 Stepping back a bit
Let’s take some time to look at a few results coming out of an earlier tradition, not strictly Principles & Parameters (and not covered by White) but still suggesting that to a certain extent L2 learners may know something (perhaps unconsciously) about “what Language is like” (which is a certain way we might characterize the content of UG).
Typological universals (Typological) universals
1960’s and 1970’s saw a lot of activity aimed at identifying language universals, All languages have vowels. properties of Language. If a language has VSO as its basic word order, Class of possible languages is smaller than then it has prepositions (vs. postpositions). you might think. If a language has one property (A), it will VSO? Yes No necessarily have another (B). Adposition type +A+B, –A–B, –A+B but never +A–B. Prepositions Welsh English Postpositions None Japanese
Markedness Markedness
Having duals implies having plurals Having plurals says nothing about having duals. “Markedness” actually has been used in a couple of different ways, although Having duals is marked—infrequent, more complex. they share a common core. Having plurals is (relative to having duals) unmarked. Marked: More unlikely, in some sense. Generally markedness is in terms of comparable dimensions, but you could also say that being VSO is Unmarked: More likely, in some sense. marked relative to having prepositions. You have to “mark” something marked; unmarked is what you get if you don’t say anything extra.
13 “Unlikeliness” Berlin & Kay 1969: Color terms
Typological / crosslinguistic infrequency. (On the boundaries of psychophysics, VOS word order is marked. linguistics, anthropology, and with issues More complex constructions. about its interpretation, but still…) [ts] is more marked than [t]. Basic color terms across languages. The non-default setting of a parameter. It turns out that languages differ in how Non-null subjects? many color terms count as basic. (blueish, salmon-colored, crimson, blond, … are not Language-specific/idiosyncratic features. basic). Vs. UG/universal features…?
Berlin & Kay 1969: Color terms Berlin & Kay 1969: Color terms
The segmentation of experience by speech symbols is essentially Arabic (Lebanon) Japanese (Japan) arbitrary. The different sets of words for color in various Bulgarian (Bulgaria) Korean (Korea) languages are perhaps the best ready evidence for such Catalan (Spain) Pomo (California) essential arbitrariness. For example, in a high percentage Cantonese (China) Spanish (Mexico) of African languages, there are only three “color words,” Mandarin (China) Swahili (East Africa) corresponding to our white, black, red, which nevertheless Tagalog (Philippines) divide up the entire spectrum. In the Tarahumara English (US) Thai (Thailand) language of Mexico, there are five basic color words, and Hebrew (Israel) here “blue” and “green” are subsumed under a single Hungarian (Hungary) Tzeltal (Southern Mexico) term. Ibibo (Nigeria) Urdu (India) Vietnamese (Vietnam) Eugene Nida (1959) Indonesian (Indonesia)
Eleven possible basic color Color hierarchy terms White, black White, black, red, green, yellow, blue, brown, Red purple, pink, orange, gray. Green, yellow All languages contain term for white and black. Blue Has 3 terms, contains a term for red. Brown Purple, pink, orange, gray Has 4 terms, contains green or yellow. Has 5 terms, contains both green and yellow. Even assuming these 11 basic color terms, there should Has 6 terms, contains blue. be 2048 possible sets—but only 22 (1%) are attested. Has 7 terms, contains brown. Has 8 or more terms, chosen from {purple, pink, orange, gray}
14 Color terms Color terms
BW Jalé (New Guinea) ‘brilliant’ vs. ‘dull’ BWR Tiv (Nigeria), Australian aboriginals in Interesting questions abound, including Seven Rivers District, Queensland. why this order, why these eleven—and there BWRG Ibibo (Nigeria), Hanunóo (Philippines) BWRY Ibo (Nigeria), Fitzroy River people (Queensland) are potential reasons for it that can be BWRYG Tzeltal (Mexico), Daza (eastern Nigeria) drawn from the perception of color spaces BWRYGU Plains Tamil (South India), Nupe (Nigeria), Mandarin? which we will not attempt here. BWRYGUO Nez Perce (Washington), Malayalam (southern India) The point is: This is a fact about Language: If you have a basic color term for blue, you also have basic color terms for black, white, red, green, and yellow.
Implicational hierarchy L2A?
This is a ranking of markedness or an implicational Our overarching theme: hierarchy. How much is L2/IL like a L1? Having blue is more marked than having (any or Do L2/IL languages obey the language all of) yellow, green, red, white, and black. universals that hold of native languages? Having green is more marked than having red… This question is slightly less theory-laden Like a set of implicational universals… than the questions we were asking about Blue implies yellow Brown implies blue principles and parameters, although it’s Blue implies green Pink implies brown similar… Yellow or green imply red Orange implies brown Red implies black Gray implies brown To my knowledge nobody has studied L2 Red implies white Purple implies brown acquisitions of color terms…
Eckman, Moravcsik, Wirth Question formation (1989) Declarative: John will buy coffee. L1: Korean (4), Japanese (6), Turkish (4) L2: English
Wh-inversion: What will John buy? Wh-fronting: What will John buy? Note L1s chosen because they are neither/neither type languages, to avoid Yes/No-inversion: Will John buy coffee? questions of transfer. Subjects tried to determine what was Greenberg (1963): going on in a scene by asking questions. Wh-inversion implies Wh-fronting. Yes/No-inversion implies Wh-inversion.
15 Eckman, Moravcsik, Wirth Eckman, Moravcsik, Wirth (1989) (1989) Example Y/N Qs: Example Wh-Qs: Did she finished two bottle wine? Why Sue didn’t look solution for her Is Lou and Patty known each other? problem? Sue does drink orange juice? Where Sue is living? Her parents are rich? Why did Sue stops drinking? Is this story is chronological in a order? Why is Patty’s going robbing the bank? Does Joan has a husband? What they are radicals? Yesterday is Sue did drink two bottles of What Sue and Patty connection? wine? Why she was angry?
% Whinv % Whfr % YNinv % WHinv Eckman et SM K 25 NO 100 YES SM K 8 NO 25 NO UA T 54 NO 100 YES Eckman et MK K 38 NO 80 NO al. (1989) TS J 70 NO 100 YES YK J 51 NO 100 YES MK K 80 NO 100 YES al. (1989) TS J 67 NO 70 NO RO J 88 NO 100 YES TM K 83 NO 100 YES wh-inv→ KO J 95 YES 100 YES RO J 85 NO 88 NO MH J 95 YES 100 YES YN-inv.→ BG T 86 NO 100 YES wh- NE T 95 YES 100 YES MA T 88 NO 100 YES SI J 95 YES 100 YES wh-inv.? UA T 91 YES 54 NO fronting? G T 100 YES 100 YES KO J 93 YES 95 YES MA T 100 YES 100 YES MH J 95 YES 95 YES ST J 100 YES 100 YES results NE T 100 YES 95 YES results TM K 100 YES 100 YES SI J 100 YES 95 YES YK J 100 YES 100 YES ST J 100 YES 100 YES
Eckman, Moravcsik, Wirth Eckman’s Markedness (1989) Differential Hypothesis
Yes/no inversion Yes (VS) No (SV) Markedness. A phenomenon or structure X in some Wh-inversion language is relatively more marked than some other phenomenon or structure Y if cross-linguistically the Yes (VS) 5 4 presence of X in a language implies the presence of Y, but the presence of Y does not imply the presence of X.
No (SV) 1 4 Duals imply plurals. Wh-inversion implies wh-fronting. Blue implies red.
(…but what counts as a “phenomenon or structure”?)
16 Markedness Differential MDH example: Hypothesis Word-final segments MDH: The areas of difficulty that a second language learner will have can be predicted on the basis of a comparison of the NL and TL such that: Voiced obstruents most marked Surge Those areas of the TL that are different from the NL and are Voiceless obstruents Coke relatively more marked than in the NL will be difficult; Sonorant consonants Mountain The degree of difficulty associated with those aspects of the TL that are different and more marked than in the NL corresponds to the Vowels least marked Coffee relative degree of markedness associated with those aspects; Those areas of the TL that are different than the NL but are not relatively more marked than in the NL will not be difficult. All Ls allow vowels word-finally—some only allow Notice that this is assuming conscious effort again. Perhaps vowels. Some (e.g., Mandarin, Japanese) allow only it need not, depending on how you interpret “difficulty” vowels and sonorants. Some (e.g., Polish) allow vowels, but it seems like Eckman means it this way. sonorants, but only voiceless obstruents. English allows all Another possible way to look at it is in terms of parameter four types. settings and (Subset Principle compliant) defaults, coupled with a FT/FA type theory…
MDH example: Eckman (1981) Word-final segments
Spanish L1 Mandarin L1 Voiced obstruents most marked Surge Voiceless obstruents Coke
Gloss IL form Gloss IL form Sonorant consonants Mountain
c
Bob [b p] Tag [tæg e ] Vowels least marked Coffee
e
Bobby [b c bi] And [ænd ] e Idea: Mandarin has neither voiceless nor voiced obstruents in
Red [rεt] Wet [w t] the L1—using a voiceless obstruent in place of a TL voiced
Wet [w te ] Deck [dεk] obstruent is still not L1 compliant and is a big markedness e jump. Adding a vowel is L1 compliant. Spanish has voiceless Sick [sIk] Letter [lεt r] obstruents, to using a voiceless obstruent for a TL voiced obstruent is L1 compliant. Bleeding [blidIn]
MDH and IL MDH and IL
The MDH presupposes that the IL obeys IL obeys implicational universals. the implicational universals too. That is, we know that IL is a language. Eckman et al. (1989) suggests that this is at So, we know that languages are such that having least reasonable. word-final voiceless obstruents implies that you also have word-final sonorant consonants, among other The MDH suggests that there is a natural things. order of L2A along a markedness scale What would happen if we taught Japanese L2 (stepping to the next level of markedness is learners of English only—and at the outset—voiced easiest). obstruents? Let’s consider what it means that an IL obeys implicational universals…
17 Generalizing with markedness Nifty! scales
Voiced obstruents most marked Surge Does it work? Does it help? Voiceless obstruents Coke Answers seem to be: Sonorant consonants Mountain Yes, it seems to at least sort of work. Vowels least marked Coffee Maybe it helps.
Japanese learner of English will have an easier time at Learning a marked structure is harder. So, if each step learning voiceless obstruents and then voiced you learn a marked structure, you can obstruents. automatically generalize to the less marked But—if taught voiced obstruents immediately, the fact that the IL obeys implicational (markedness) universals structures, but was it faster than learning means that voiceless obstruents “come for free.” the easier steps in succession would have been?
The Noun Phrase Accessibility The Noun Phrase Accessibility Hierarchy Hierarchy
Keenan & Comrie (1977) observed a hierarchy among the kinds of There are several kinds of relative clauses, based on relative clauses that languages allow. where the head noun “comes from” in the The astronaut [(that) I met yesterday]. modifying clause: Head noun: astronaut The astronaut… Modifying clause: (that/who) I met — yesterday. [I met — yesterday] object Compare: I met the astronaut yesterday. [who — met me yesterday] subject This is an object relative because the place where the head noun would be [I gave a book to —] indirect object in the simple sentence version is the object. [I was talking about —] obj. of P [whose house I like —] Genitive (possessor) [I am braver than —] obj. of comparative
The Noun Phrase Accessibility Resumptive pronouns Hierarchy
Turns out: Languages differ in what positions The guy who they don’t know whether he wants they allow relative clauses to be formed on. to come. A student who I can’t make any sense out of the papers he writes. English allows all the positions mentioned to be used to make relative clauses. The actress who Tom wondered whether her father was rich. Arabic allows relative clauses to be formed only with subjects. Greek allows relative clauses to be formed only In cases where relative clause formation is not with subjects or objects. allowed, it can sometimes be salvaged by means of a pronoun in the position that the head noun is to be associated with.
18 NPAH and resumptive NPAH pronouns
Generally speaking, it turns out that in languages which do not allow The positions off which you can relativize relative clauses to be formed off a certain position, they will instead allow relative clauses with a resumptive pronoun in that position. appears to be an implicational hierarchy.
Arabic: allows only subject relative clauses. But for all other positions allows a resumptive pronoun construction, analogous to: The book that John bought it. Lang. SUB DO IO OP GEN OCOMP The tree that John is standing by it. Arabic – + + + + + The astronaut that John gave him a present. Greek – – +? +? + + Japanese – – – – +/ – Persian – (+) + + + +
Noun Phrase Accessibility Noun Phrase Accessibility Hierarchy Hierarchy More generally, there seems to be a More generally, there seems to be a hierarchy of “difficulty” (or hierarchy of “difficulty” (or “(in)accessibility”) in the types of relative “(in)accessibility”) in the types of relative clauses. clauses. A language which allows this… A language which allows this… Will also allow these.
Subj > Obj > IO > OPrep > Poss > OComp Subj > Obj > IO > OPrep > Poss > OComp
Noun Phrase Accessibility Relation to L2A? Hierarchy More generally, there seems to be a Suppose that KoL includes where the target hierarchy of “difficulty” (or language is on the NPAH. “(in)accessibility”) in the types of relative Do L2’ers learn the easy/unmarked/simple clauses. relative clauses before the others? Do L2’ers transfer the position of their L1 first? A language which allows this… Does a L2’ers interlanguage grammar obey this Will also allow these. But not these… typological generalization (if they can relativize a particular point on the NPAH, can they relativize everything higher too?)? Subj > Obj > IO > OPrep > Poss > OComp
19 NPAH and L2A? NPAH in L2A Probably: The higher something is on the NPAH, the easier (faster) it is to learn. Very widely studied implicational So, it might be easier to start by teaching subject relatives, universal in L2A—many people have then object, then indirect object, etc. At each step, the difficulty would be low. addressed the question of whether the IL But, it might be more efficient to teach the (hard) object of a obeys the NPAH and whether teaching aa comparison—because if L2’ers interlanguage grammar marked structure can help. includes whatever the NPAH describes, knowing that OCOMP is possible implies that everything (higher) on the NPAH is possible too. That is, they might know it without Eckman et al. (1989) was about this second instruction. (Same issue as before with the phonology) question…
Change from pre- to post-test Transfer, markedness, … Eckman, Bell, & Nelson (1988) Do (2002) looked at the NPAH going the other way, English→Korean. English: Relativizes on all 6 positions. Korean: Relativizes on 5 (not OCOMP)
S SU do IO OP GE 13 + + + + + 14 + + + + - 16 + + + - - 29 + + - - - 31 + - - - - 20 - - - - -
Transfer, markedness, … Subset principle? A tempting analogy… in some The original question Do was looking at cases, parameters seem to be was: Do English speakers transfer their I ranked in terms of how position on the NPAH to the IL Korean? E permissive each setting is. But look: If English allows all 6 positions, why do some of the learners only Null subject parameter relativize down to DO, some to IO, some Option (a): Null subjects are permitted. to OPREP? Option (b): Null subjects are not permitted. It looks like they started over. Italian = option a, English = option b.
20 Reminder: Subset Principle Reminder: Subset Principle
The idea is If one has only positive evidence, and I If parameters are organized in terms of The Subset Principle is basically that learners permissiveness, are conservative—they only assume a grammar Then for a parameter setting to be learnable, the E starting point needs to be the subset setting of the sufficient to generate the sentences they hear, parameter. allowing positive evidence to serve to move them to a different parameter setting. The Subset principle says that learners should Applied to L2: Given a choice, the L2’er start with the English setting of the null subject parameter and move to the Italian setting if assumes a grammatical option that generates a evidence appears. subset of the what the alternative generates. Does this describe L2A? Is this a useful sense of markedness?
Subset principle and Subset vs. Transfer markedness Based on the Subset principle, we’d expect the The Subset Principle, if it operating, would say that L2A unmarked values (in a UG where languages are starts with all of the defaults, the maximally conservative learnable) to be the ones which produce the “smallest” grammar. grammars. Another, mutually exclusive possibility (parameter by parameter, anyway) is that L2A starts with the L1 setting. Given that in L1A we don’t seem to see any “misset” This means that for certain pairs of L1 and L2, where the L1 has parameters, we have at least indirect evidence that the the marked (superset) value and L2 has the unmarked (subset) value, only negative evidence could move the L2’er to the right Subset principle is at work. Is there any evidence for it in setting. L2A? Do these NPAH results constitute such evidence? Or, some mixture of the two in different areas.
NPAH and processing? NPAH and processing?
At least a plausible alternative to the NPAH results If it’s about processing, then the reason following from the Subset Principle is just that L2’ers progress through the “hierarchy” relative clauses formed on positions lower in the hierarchy are harder to process. Consider: might be that initially they have limited The astronaut… processing room—they’re working too hard
who [IP t met me yesterday] SUB at the L2 to be able to process such deep who [IP I [VP met t yesterday]] DO extractions. who [IP I [VP gave a book [PP to t ]]] IO Why are they working so hard? who [IP I was [VP talking [PP about t ]]] OPREP
whose house [IP I [VP like [DP t ’s house]]] GEN (Well, maybe L2A is like learning history?) who [IP I am [AP brave [degP -er [thanP than t ]]]] OCOMP
21 NPAH and processing? Subset problems?
Is the NPAH itself simply a result of processing? One problem, though, is that many of the The NPAH is a typological generalization about parameters of variation we think of today don’t languages not about the course of acquisition. seem to be really in a subset-superset relation. So there has to be something else going on in these Does Arabic have a lower threshhold for processing difficulty than English? Doubtful. cases anyway. V→T The NPAH may still be real, still be a Yes: √ SVAO, *SAVO markedness hierarchy based in something No: *SVAO, √ SAVO grammatical, but it turns out to be confounded by Anaphor type processing. Monomorphemic: √ LD, *Non-subject So finding evidence of NPAH position transfer is Polymorphemic: *LD, √ Non-subject very difficult.
22