Vallduví & Vilkuna (1998) On Rheme and Kontrast Syntax an Semantics 29

Perspectives on Focus (2) Contrast, rheme, information, exhaustiveness, … Vallduví & Vilkuna (1998) “On Rheme and Kontrast” Syntax an Semantics 29 John introduced BILL to Sue. Some options for what kind of kontrast we see here: (3) ☞ if M = {a, b, c} and P(x ∈ M), … a. Identificational kontrast: …then P(a). Consider rheme and kontrast, both of which are concepts commonly lumped under focus: Rheme: The new information of the sentence. The elements in the sentence that are contextually unbound. What is asserted rather than presupposed. Kontrast: An “operator-like” element that’s contrastive, exhaustive, identificational …or something… M = {Bill, Mary, Pat}; P = being introduced to Sue by John; if P is true of some x in M, then P is true of Bill. Assuming John introduced somebody to Sue, John introduced Bill to Sue. b. Point: These are not two aspects of a unified focus—they are two different things. Exhaustive kontrast? …then ¬ ( P( (y ≠ a) ∈ M) ) …then P is not true of any y ∈ M other than a. V&V assert that (3b) is too strong, end of story. Sounds directly contrary to É. Kiss (1998) (“identificational focus is exhaustive”). Might be nice to hear arguments. Re: Rheme: Rheme: The (context) “update potential” of the utterance. i.e. how this utterance is to change your view of the world. (if a theme is provided, it tells you where to make that change) (1) a. a----What about the pipes? What’s wrong with them? The pipes are [R RUSTY ] b. a-----------Why does the water from the tap come out brown? 1a-------- I have some rust remover. You have anything rusty? [R [R The PIPES ] are rusty ] c. a-------------I wonder whether the pipes are rusty. The pipes [R ARE ] rusty. The information state (context) contains familiar entities and conditions. e.g. rusty(the pipes) ‘the pipes are rusty’, λx.rusty(x) ‘something is rusty.’ λP.P(the pipes) ‘something (salient?) is true of the pipes.’ Theme zeroes in on one of those conditions, Rheme provides the missing information. Koncerning Kontrast: If a is kontrastive, a MEMBERSHIP SET M = {…, a, … } “is generated” and “becomes available to semantic computation as some sort of quantificational domain.” Fighting For The Kause: Theme–Rheme is not equal to Kontrastive–Nonkontrastive Kontrast is orthogonal to theme–rheme. Look: They can overlap: (4) a. only requires a kontrast (explicitly exhaustive) What did you see? I saw only [R the DOG ]. the dog is new information But they can also not overlap: b. Why are you so excited? [R There’s only a MONTH till Christmas now ]. the whole thing is new information only still requires a kontrast “Unless a distinction between kontrast and rheme is made, examples like (4b) can be accounted for only if one allows for the possibility of having foci-within-foci.” Well, why not do that? Could we handle these (and other) cases without recourse to two different concepts if we allow something like recursive focus-ground structure? Has it been done? Has it been argued for? Things to keep in mind… People like to think of wh-expressions as some subkind of focus. Why? Well… Still not konvinced? Ok, let’s go to syntax… Hungarian: More generally: Crosslinguistic evidence for the syntactic reality of [±Rh] and [±K]. Wh-phrases and foci seem to land in the same structural position. Wh-answers usually have focus where the wh-word was. Well, focus comes in two kinds, kontrast and rheme. And wh-phrases can’t be rhemes— after all, in the question, the wh-word isn’t necessarily the rheme. not the answer, I’m not talking about the answer, I’m talking about the question. (5) I don’t care about who Bill introduced to Sue… Who did JOHN introduce to Sue? At the very least, John is new information here. Maybe even the rest (including who?) is old information. Claim : (9) K-field [+K] [CP T-field [–K,–Rh] [IP What things did Anna get for her birthday? What is it that Anna got for her birthday? What about the flowers? Did Anna have to buy them or did she get them for free? Interrogative kontrast: If M = {a, b, c} and P(x ∈ M) then ?x P(x). Testing the typology—if kontrast and rhematicity are orthogonal then you should be able to find (a) nonkontrastive themes (regular topics) and (b) kontrastive themes (“contrastive topics”). And, of course, you can. Thematic kontrast: if M = {a, b, c} and P(a) then P′((y∈M) ≠ a) ‘if P is true of a then other properties like P′ hold of the non-a members of M.’ (8) How’d she do at the race? (Well,) the first 100 meters, she ran in RECORD TIME. (…but after that, she fell behind) i.e. P = being run in record time by her, a = the first 100 meters, M={a, the rest of the race}, P′ = being run in an unsuccessful manner by her. About only: only has to associate with kontrasts. (and, they assert, it can do it even in the theme. Maybe Japanese dake-wa is evidence). V-field [+Rh] [I′/VP Finnish Anna sai [RKUKKIA]. [R[KKUKKIA]] Anna sai. [KKukkia] Anna [RSAI]. +K –K, –Rh +Rh Hungarian: “focus moves to preverbal position”—but what kind of focus? • an “identificational” operator, an “exhaustiveness” operator • so it’s kontrast, right? • rhemes stay in situ (sounds like “information focus” doesn’t it?) (10) Where did Peter put the book? (idea: rheme will be on the answer) Péter letette a könyvet [R a POLCRA ]. Peter down-put the book the shelf-on ‘Peter put down the book on the shelf’ I have no idea what this means, but the idea is just that M is the same set that gives you alternatives to the wh-word in what counts as an answer to the question. (7) [+K] elements (contrastive focus, contrastive topic) move to CP. Noncontrastive “main news” (rheme) stays low, below IP Topics (theme) in between. Interrogative wh-words are kontrasts—not rhemes. Though I’d bet that people who called wh-words “focus” were thinking of something like kontrast anyway—I think they’re generally people who don’t consider theme–rheme. (6) Finnish: vs. (11) Where-all did Peter put the book? Péter [R a POLCRA ] letette a könyvet. Peter the shelf-on down-put the book ‘It’s on the shelf (and nowhere else) that Peter put down the book’ But topichood ([–Rh]) takes priority over kontrastiveness ([+K]) unlike in Finnish. (12) Topic [–Rh] [TopP Kontrast [+K] [FocP Rheme [+Rh] [IP/VP Catalan: [–Rh] causes things to leave the IP (recall, links go left, tails go right) (13) El Joan va deixar una nota damunt la TAULA. El Joan hi2 va deixar una NOTA, damunt la taula2. El Joan l3’hi2 va DEIXAR, una nota3, damunt la taula2. ‘John left a note on the table.’ a. b. c. Hungarian (13a) is typical—John = theme, table et al. = rheme In (13c), all of the arguments are part of the theme, e.g. “—John did something with a note and the table… —(13c)”. There, we’ve now derived that wh-words and kontrasts should act alike. Except that they don’t: • wh-words have to be directly adjacent to the verb • focus fronted stuff can have things between the focus and the verb. Why? • In questions, wh-word is only part of the rheme (recall: rhemes are IP-internal) (13′) (15) In (13b), on the table is part of the theme, e.g. “—After lunch I laid out all the gifts on the table. —Oh, by the way, did you see? (13b)”. a. b. c. [El Joan]Topic [va deixar una nota damunt la TAULA]Focus. [El Joan]Topic [hi2 va deixar una NOTA]Focus, [damunt la taula2]Tail. [El Joan]Topic [l3’hi2 va DEIXAR]Focus, [una nota3]Tail, [damunt la taula2]Tail. They propose: [–Rh] dislocate out of the clause. That which remains in the clause is [+Rh]. So: [–Rh] matters to Catalan, like [+K] mattered to Finnish. (and [–Rh] doesn’t suffice in Catalan to tell whether it right- or left-dislocates; that’s a third distinction [±link] if you like.) but does [+K] have any effect in Catalan? [R On va ficar el LLIBRE, ] la Maria? where put the book, the Maria ‘Where did Mary put the book?’ • In the “focus-fronted” cases, the kontrast is the lone rheme element. (16) a. a--------l 1 a--l1a--------------l ? ? 111 ? Kontrast fronting—is it really? Hint: no. b. There is this thing called “focus fronting” (kontrast fronting) wherein the focused element shows up left of the verb. Any other language, we’d say “sure, it moves there. You know, like wh-movement. Or like Hungarian…” Ah, but it’s an illusion— (14) [CP/XP … non-quantifiers … [IP quantifiers … I°–V° … ] ] So, if you see a preverbal non-quantifier, it’s outside of IP, and must have gotten there by being non-rhematic. So, wh-words move to SpecIP, making them strictly preverbal. Should be the same for negative quantifiers, other quantifiers. And other kontrasts. (maybe quantifiers = things that deal with an alternative set?) [IP un ROLEXi [I tj ] [VP tj ti tk ] ]— s’ha compratj — el Jordik. 1 " z--------------m (kontrastive rheme) a------la------------l ? 11 (and we’re looking for it to be an illusion because we want to believe Vallduví that rhemes are all IP-internal—but, if there is focus movement of a rhematic element to an FP outside of IP, we can’t) Is it like wh-movement? Well, even if it is (which it isn’t), wh-movement itself doesn’t escape IP. (according to a previous argument “on the basis of ample evidence” (oh, ok…) that puts them in SpecIP, which is specialized for hosting “quantificational elements” including wh-words, negative quantifiers, other quantifiers. Non-quantificational preverbal arguments are further left, outside of NP.) [R Un ROLEX ] s’ha comprat el Jordi. a Rolex self.has bought the Jordi ‘A ROLEX John bought himself.’ b. [IP ? [I tj ] [VP tj un ROLEX tk ] ]— s’ha compratj — el Jordik. 1 " z--------------m (nonkontrastive rheme) (That’s a little weird, postposing the verb alone like that… why couldn’t it just be a remnant? i.e. front un ROLEX to a position outside VP but inside IP, thematize VP when nothing’s left in it but the verb.) Bottom line: [+K] in Catalan does have a structural effect, it’s just more subtle than in Finnish. ([+K] moves to SpecIP, [–Rh] moves out of IP). If you like to think of things this way: • Displace-[–Rh] >> Front-[+K] • Front-[+K] >> Displace-[–Rh] (17) The flowers I gave only to John. –Rh +K in Catalan and Hungarian in Finnish • Displace-[–Rh] >> syntax >> Front-[+K] in English (topicalization anyway) Ponderables… Recall, Vallduví in his solo career distinguishes focus (=rheme) from ground (=theme), and in ground, he distinguishes between links and tails. Fn. 3 claims that the link–tail distinction is NOT important to this paper, which must mean that whatever differentiates links from tails is not what differentiates kontrasts from nonkontrasts. Fn. 4 tells us that only links can be “contrastive topics”—tails can’t be (they’re not, in fact, “topics” at all). But can there be a contrastive tail? What is the correlate of É. Kiss’ (1998) “identificational vs. information focus” in V&V? • identificational focus = kontrast, the kind that moves. —V&V assert that kontrast need not be exhaustive. —K says identificational focus need not be contrastive (!?). (so is her “contrastive” the same as V&V’s “kontrastive”?) • Is information focus = rheme? I think so.