Center for the Study of Language and Information Report No. CSLI-87-112

October 1987

Bare Plurals, Naked Relatives, and Their K in

D i e t m a r Zaefferer University of Munich and CSLI, Stanford University

C S L I was founded early in 1983 by researchers from Stanford University, S R I International, and Xerox P A R C to further research and development of integrated theories of language, information, and computation. C S L I head­quarters and the publication offices are located at the Stanford site.

C S L I / S R J Internat ional C S L I / S t a n f o r d C S L I / X e r o x P A R C 333 Ravens wood Avenue Ventura Hall 3333 Coyote Hill Road Menlo Park, C A 94025 Stanford, C A 94305 Palo Alto, C A 94304

Bare Plurals , Naked Relatives, and Their K i n Towards a Unified Semantics of German wh -Constructions

Dietmar Zaefferer*

1. I N T R O D U C T I O N

1.1. Terminological clarification

Free relative clauses - or naked relatives, as I w i l l call them for reasons that w i l l become clear later - are members of a fami ly of related, but dist inct constructions, which are often referred to by the name of wh-constructions, because they are marked in Engl ish by the presence of wh-words l ike who or where. The t e rm can be carried over to any language w i t h interrogative words and can be conceived of as denoting the fami ly of construction types that have the interrogative words and their homonyms as common denominator. Therefore, i f there are no homonyms, the wh-constructions are just the constituent interrogatives. Most languages, however, do have homonyms and so e.g. both i n Engl ish and i n German, the headless or free relative constructions are also i n the fami ly of wh-constructions (although, i n this sense, i n German the noun-headed relatives are not).

1.2. Philosophical confusion

According to Eike von Savigny ([9]), the interrogative/relative homonymy can even lead to philosophical confusion. He points out that i n his "Treatise Concerning H u m a n Understanding" , John Locke uses a sentence w i t h a wh-construction that sounds irrefutable as evidence for the position he advocates (i.e. the certainty of inter ior perception; [6]vol.11:269):

(1). . . it being impossible but that he should perceive what he perceives...

A closer inspection, however, shows that i t is unassailable only i n its free relative reading, which is, i n fact, tautological and therefore not very helpful for Locke's point. The other, interrogative, reading, however, the one Locke needs for his argument, is far from being

2

obvious. I n order to point out the cr i t i ca l semantic difference, I restate the relevant portion of (1) and give a (tentative) logical representation for each reading:

(2) John perceives what he perceives.

(2a) \fx[perceive'(jtx)-> perceive*(j,x)] (free relative reading)

(2b) \fxiperceive*(j,whether%[perceive'(j,x)])] ( interrogative reading)

(2a) shows the i r re futab i l i ty of the free relative reading: Whenever John perceives something, he perceives i t . True, but uninterest ing. The interrogative reading is b u i l t on the assumption that a wh-interrogative is but a generalization of a whether-interrogative. I t says that for any object x, John perceives whether or not he perceives x. And i t entails that i f John perceives x, he also perceives whether he perceives x. That is an interesting c laim (note its impl i c i t recursiveness), but far from being unassailable.

1.3. The relative/interrogative distinction in German

The Locke example was the one through which I got f irst interested i n the free relative/ interrogative distinction and i t was easy to see that the same dist inction holds for German as wel l . To i l lustrate this , I w i l l use a somewhat t r i ck ier example. Its perplexing apparence is due to the fact that the two readings are truthcondit ional ly independent, which makes i t possible to construct examples where a sentence expression is consistently conjoined w i t h its negative counterpart as i n (3):

(3) Eva weiss, was Max wissen will, aber was er wissen will, weiss sie nicht.

(Eva knows what Max wants to know, but she does not know what he wants to know.)

This sounds mind-boggling at f i rs t , but a moment's reflection shows that i t is possible to give i t a sensible interpretat ion. I f the first occurrence of was Max wissen will is read as a p la in interrogative, and the second one as a relative one, then (3) means something l ike Eva knows the issue Max is interested in, but she doesn't know the answer and i t is easy to t h i n k of circumstances which would make (3) true: Suppose that (a) Eva doesn't know whether she is happy, (b) the only t h i n g Max wants to know is whether Eva is happy, and (c) Eva knows about Max ' wish. It 's of course awkward but possible to express this , abstracting from the concrete instant iat ion , as (3).

3

This gives a clear indicat ion of what the two readings of was Max wissen will mean: As a p la in interrogative, i t means roughly a collection of propositions that are denoted by the sentences that can be obtained form the open sentence M a x w i l l χ w issen by proper instant iat ion for x, as a relative, i t means a collection of propositions χ which satisfy this open sentence. A sketch of a logical representation of the discussed reading of (3) would therefore be (3a):

(3a) \/x[wiss'(etwhether[wiss-woir(m,x)]) &

~~1 \fx[wiss'-woll\m,x) --> wiss'(e,x)]

4

2. BARE A N D DRESSED, R E L A T I V E S A N D P L U R A L S

2.1. The endofa nice analysis

Our short discussion of the interrogative/relative ambigui ty of some wh-constructions involved a generalizing analysis of the meaning of free relatives which has the nice property of g iv ing an straightforward account of inferences l ikes the fol lowing:

(4) Ευα weiss, was Max wissen will. [relative reading]

(Eva knows what Max wants to know.)

(5) Max will wissen, ob es regnet.

(Max wants to know whether it is raining.)

Therefore:

(6) Ευα weiss, ob es regnet.

(Eva knows whether it is raining.)

This analysis, however, happens to coincide w i t h the one which is most natura l ly given for universal ly quantif ied relatives l ike alles, was Max wissen will (everything Max wants to know), which i n inferences l ike the one above, behave exactly al ike. I f their semantic behaviour could be shown to be identical , the synonymy assumption expressed i n the identical analyses would be val id . Unfortunately , there are at least three kinds of counterexamples to this assumption. The first k i n d was brought to my attent ion by Irene Heim (p.c.), and has to do w i t h the so-called adverbs of quanti f icat ion (cf. Lewis[4]).

2.1.1 Counterevidence A: Adverbs of quantification

These examples show that expl ic i t ly universal ly quanti f ied relatives (u-relatives for short) cannot be freely interchanged w i t h free relatives i n the presence of adverbs of quanti f ication.

(7) Was Eva am Strand findet, ist meistens wertlos.

(What Eva finds at the beach is mostly worthless.)

5

(8) * Alles, was Ευα am Strand findet, ist meistens wertlos.

(^Everything that Ευα finds at the beach is mostly worthless.)

What is going on here? A n i n t u i t i v e way of describing i t is the following: The impl i c i t generalization felt i n the free relative can be overridden by the explicit most quanti f ier , whereas the expl ic it universal quanti f icat ion of the u-relative clashes w i t h the most quanti f icat ion of the adverb, y ie ld ing a semantically anomalous sentence.

( I f meistens is read as a temporal adverb, (8) is perfectly normal , but that reading has a quite different meaning (suggesting that Eva finds, say, Christmas trees, which are worthless except once a year) and should not distract us.)

People w i t h the PTQ-truth def init ion (cf. Montague[7]) i n mind might come up w i t h the fol lowing proposal for a fix: Let free relatives be represented as free variables, w i t h the same variable being free in the rest of the sentence. I f no adverb of quanti f icat ion intervenes, then the whole sentence is true (according to the PTQ definition) i f i t is true for a l l variable assignments, i.e. i f the u-quantified counterpart is true as wel l . Adverbs of quanti f icat ion, however, b ind the free variable, y ie ld ing something l ike (7a) as meaning representation for (7), and cannot find anyth ing to b ind i n case (8).

(7a) MOST x[am-Strand-find(e,x): wertlos\x)]

This looks nice, but i t s t i l l doesn't work, since i t would always assign widest scope to the i m p l i c i t universal quanti f ier of the free relative, and this is i n clear contradiction w i t h data l ike the fo l lowing, where the narrow scope reading is the preferred one, or even the only one:

(9) Ein Polizist υerhaftete, wen er im Hause antraf.

(A policeman arrested whoever he met in the house.)

So what we need is some k i n d of device which de l imits the scope of the cr i t ical variable, but does not specify any k i n d of quanti f icat ion u n t i l an adverb of quanti f icat ion is met, which then determines the k i n d of quanti f i cat ion, or else universal quanti f icat ion is filled i n as a default. But before we t ry to spell this out, i t m ight be useful to have a look at some more countervidence to the synonymy assumption.

2.1.2 Counterevidence B: Degree particles

6

A second class of counterexamples was brought to my attent ion by Bertold Brecht (posthumous communication), who wrote the fine line (10):

(10) Nur wer im Wohlstand lebt, lebt angenehm.

{Only he who lives in wealth, lives a pleasant life.)

Again , subst i tut ing a u-relative for the free relative results i n a semantically anomalous sentence:

(11) *Nur jeder der im Wohlstand lebt, lebt angenehm.

(*Only everybody who lives in wealth, lives a pleasant life.)

(There are acceptable readings for (11) where the focus of nur is only on some part of the NP, but these shall be ignored here.)

Basically the same holds for other degree particles l ike gerade, sogar, auch, although the degree of deviance of the u-relative counterpart may be smaller.

(12) Gerade/Sogar/Auch (? alles) was ich nicht wissen soll, möchte ich wissen.

(I want to know precisely/even/also (? everything that/what) lam not

supposed to know)

Interest ingly enough, adding a nur (only) to a free relative does not simply add the inverse impl icat ion to the or ig ina l one, but seems to replace i t . Cf. (13) - (16):

(13) Max hat Geld.

(Max has money.)

(14) Wer Geld hat, hat Macht.

(He who has money has power.)

(15) Nur wer Geld hat, hat Macht.

(Only he who has money has power.)

(16) Max hat Macht.

(Max has power.)

7

(16) can be inferred from (13) only w i t h the help of (14), not w i t h (15). This is some more evidence for the vo la t i l i ty of the generalizing power of free relatives.

2.1.3 Counterevidence C: Logical particles

The t h i r d group of counterexamples against the free relative/u-relative synonymy assumption behaves in a way which is complementary to the ones in the groups A and B. Adding logical particles l ike nicht or fast to a relative NP works fine, i f i t is a u-relative, but is out w i t h free relatives:

(17) Nicht/Fast jeder, der wagt, gewinnt.

(Not/Almost everybody who dares wins.)

(18) *Nicht/*Fast wer wagt, gewinnt.

(* Not/* Almost who dares wins.)

(As before, special stress can change the picture: The f irst var iant of (18) can be made acceptable i f i t is given contrastive stress and i f i t is followed by a correcting statement l ike "... but who pays".)

Aga in , i f free relatives would contain a hidden universal quanti f ier , one would expect i t to be modifiable i n the usual ways. But i t isn't. So the conclusion to be drawn from these data is s imply that free relatives do not contain any hidden quanti f ier and that they derive their generalizing meaning from a different source.

2.2 Bare and naked

A t least one property of free relatives, namely the presence of a strong feeling of quanti f icat ional power i n the absence of any explicit quanti f ier reminds one strongly of another construction type which was not properly understood for a long t ime , namely the so-called bare plurals (cf .Carlson[l]) . U n l i k e their equally undressed clausal relatives (henceforth naked relatives), they come i n two varieties, an existential and a generalizing one. (19) i l lustrates the former, (20) the latter:

(19) Diebe haben unser Auto ausgeräumt.

(Thieves have emptied our car.)

8

(20) Diebe sind ehrlos.

(Thieves are dishonorable.)

The existential var iant seems to have no relative clause counterpart, at least (22) is not a paraphrase of (19) , given (21):

(21) Wer stiehlt, ist ein Dieb, und umgekehrt.

(He who steals is a thief, and vice versa.)

(22) Wer stiehlt, hat unser Auto ausgeräumt.

(He who steals has emptied our car.)

But (20) does have relative clause counterparts, e.g. (23), and (23) is a paraphrase of (20), given (21):

(23) Wer stiehlt, ist ehrlos.

(He who steals is dishonorable.)

Another difference between bare plurals and naked relatives is that only the latter admit of a definite reading. Cf. (24), where a- and b-variant stand i n a paraphrase relation, and (25), where this is not the case:

(24a/b) (Was/Das, was) Hans gestern gesagt hat, stimmt nicht.

(What Hans said yesterday is not correct.)

(25a/b) (Äußerungen/Die Äußerungen), die Hans gestern gemacht hat,

stimmen nicht.

((Remarks/The remarks) that Hans made yesterday are not correct.)

H a v i n g stated the fundamental differences between bare plurals (henceforth BP's) and naked relatives (henceforth NR's), we can now proceed to a step-by-step check of the s imi lar i t ies between the two.

The first paral lel : Adverbs of quanti f icat ion are fine w i t h bare, but not w i t h universal ly quanti f ied plurals (cf. examples (7) and (8) above):

(26) Fundsachen sind meistens wertlos.

9

(Things found are mostly worthless.)

(27) *Alle Fundsachen sind meistens wertlos.

(*All things found are mostly worthless.)

The second parallel : BP's, but not u-quantif ied plurals go w i t h grade particles (cf. (10) and (11) above):

(28) Nur Reiche leben angenehm.

(Only rich people live a pleasant life.)

(29) *Nur alle Reichen leben angenehm.

(*Only all rich people live a pleasant life.)

T h i r d parallel : The same logical particles which can operate on universal quantif iers, but not on NR's (cf above (17) and (18)), are impossible w i t h BP's:

(30) Nicht/Fast alle Seeleute sind abergläubisch.

(Not/Almost all sailors are superstitious.)

(31) *Nicht/*Fast Seeleute sind abergläubisch.

(*Not/*Almost sailors are superstitious.)

(Note that the same qualif ications concerning the intended reading of the starred examples hold as above.)

F ou r t h paral le l (and the last one I want to mention here): The generalizing flavor of both of them makes them especially apt for prototypical, legal, and proverbial usage:

(32) Echte Bayern trinken Bier.

(True Bavarians drink beer.)

(33) Wer ein echter Bayer ist, trinkt Bier.

(Whoever is a true Bavarian drinks beer.)

(34) Falschparker werden abgeschleppt.

(Illegally parked cars get towed.)

10

(35) Wer falsch parkt, wird abgeschleppt.

(Whoever parks his car illegally gets towed.)

(36) Lügen haben kurze Beine.

(Lies have short legs.)

(37) Wer einmal lügt, dem glaubt man nicht, und wenn er auch

die Wahrheit spricht.

(Whoever once tells a lie will not be believed, even if he tells

the truth.)

So one desideratum for a semantics of (German) NR's is that i t account for these s imi lar i t ies w i t h , and differences from BP's. Another desideratum is of course, and this takes us back to the introductory remarks, that i t assign them the ir proper place in the family of related constructions.

11

3. TOWARDS A U N I F I E D S E M A N T I C A N A L Y S I S OF G E R M A N wh-CONSTRUCTIONS

3.1. Six types of German wh-constructions

What does the fami ly of German wh-constructions ( in my strict sense) look like? I c la im that contrary to what one might guess from the previous examples and from the l i terature , i t contains not j u s t two members, interrogative and naked relative clauses (remember that fu l ly dressed, i.e. noun-headed relatives are, unl ike most of their English counterparts, not l i t e ra l l y wh-constructions, since they contain no wh-words, the relative pronoun having i n these cases a form which starts w i t h a d), but some more, apparently even six. According to the size and the category of the wh-construction, I dist inguish 3 groups:

- Lexical wh-constructions, where wh-word and wh-construction coincide,

- Phrasal wh-constructions, where the wh-word precedes a clause to form a phrase of the same category as the wh-word (which may be, but need not be the category clause), and

- Clausal wh-constructions, where the wh-word precedes a clause to form a clause, regardless of the category of the wh-word.

N a k e d relatives are phrasal constructions; i n the wer-cases considered so far, they form personal terms (i.e. noun phrases w i t h the feature [ + person]), but there are other cases as we l l , e.g. local adverbials such as i n (38):

(38) Fritz lebt wo sich Hasen und Füchse gute Nacht sagen.

(Fritz Hues where hares and foxes say good night to one another.)

E m b e d d e d interrogatives are clausal wh-constructions, no matter i f the wh-word is an NP or an adverbial or what have you, cf. (39):

(39a/b) Eva verschweigt Max, wen sie liebt I wo sie lebt.

(Eva conceales from Max who she loves I where she lives.)

So far, so obvious. The inclusion of the fol lowing four types, however, may deserve some comment.

The least surpris ing one may be that of the pseudo-clefts, since they can be considered as a special case of free relatives (cf. [3]). (40) is an i l lus t ra t i on :

12

(40) Was Max am meisten verblüffte, war Evas Frechheit.

(What surprised Max the most was Eva's insolence.)

The question is: W h a t is i t that makes them special?

Second, w i t h lexical wh-constructions, there can be only one type, since here, the construction consists simply i n the wh-word itself. This construction type is not very much used i n w r i t t e n German, but quite frequently i n the colloquial language. Its meaning is much the same as t h a t of indefinites l ike jemand, etwas, etc., but unl ike the latter , its representatives cannot be stressed and have to occur after the finite verb i n verb-second sentences. I therefore call them weak indefinites. Here is an example.

(41) Da singt wer.

(There is somebody singing.)

T h i r d and fourth , the category of clausal wh-constructions contains, alongside w i t h the interrogat ive clauses, two l i t t l e known construction types: wh-exclamatories as i n (42), and the antecedents of what has been called irrelevance conditionals, which I w i l l cal l no-matter-conditionals , i l lustrated by (43):

(42) Es ist unglaublich, wie frech Eva ist.

(It is incredible how insolent Eva is.)

(43) Was du auch sagst, ich werde meine Entscheidung nicht ändern.

(Whatever you say, I will not change my decision.)

The wh-exclamatories (cf. [11]) look pretty much l ike wh-interrogatives (and thus have g iven rise to some confusion among grammarians) , but unl ike them, they fit into argument places where also da/?-clauses (declarative clauses) fit, but no o6-clauses ( interrogat ive clauses):

(44) Es ist unglaublich, daß I*ob Eva so frech ist.

(It is incredible that I* whether Eva is so insolent.)

More impor tant ly , their semantics differs considerably from that of interrogatives. More on t h a t below.

No-matter-conditionals are almost ignored i n the l i terature , there seems to be no established t e r m for them, and that is why I have dubbed my own term. One property sets the i r antecedents apart from a l l the other constructions considered so far: they cannot be

13

confounded w i t h the latter , since they are obl igatori ly marked w i t h an auch or an immer i n the clause, or w i t h an egal or gleich preceding i t (cf. Engl ish -ever, no matter):

(45) Egal, wo du hingehst! Wo immer du hingehst / *Wo du hingehst,

ich werde dich finden.

(No matter where you go Iwhereever you go I * where you go,

I will find you.)

Semantically, they seem to be conditionals, but since their antecedent i n general does not impose any real restrictions, the pragmatic effect is mostly jus t the opposite: Instead of re la t iv i z ing the t r u t h of the consequent to a given condition, they emphasize the unresticted va l id i ty of the consequent.

Next , I w i l l present an overview of which wh-words occur in what funct ion.

3.2. A synopsis of German wh-constructions

Table 1 gives an answer to the question: Which German wh-words occur i n which wh-constructions? The fol lowing abbreviations are used i n the table:

Heads of the columns: IC: Interrogative clause, NCA: no-matter-conditional antecedent, EC: Exclamatory clause, PC A: Pseudo-cleft argument, W I : Weak indefinite. Heads of the rows: the was and wie are subcategorized according to the category which they represent: wasNF, wasOC,

waslc, and wasw mean the NP, declarative clause, interrogative clause, and i n f i n i t i v e phrase was, respectively; wieW9rR means the instrumenta l wie as in Wie öffnet man diese Türe? (How

does one open this door?), wieU0OE the modal wie as i n Wie singt Eva? (How does Eva sing?), and u ; i e S P E C the preadjectival/preadverbial specifier wie as i n Wie groß ist Max? (How tall is Max?)).

A plus sign means that grammatical examples can be found, i t does not i m p l y that they are often used. Question marks are placed where only questionable examples can be given, minus signs say that a l l examples are out. The entry i n row 9, co lumn d. indicates that a l though there are no relative temporal phrases w i t h wann, the closely related wenn is possible i n this construction.

14

Table 1

Clausal Phrasal Lexical

IC NC A EC NR PCA W I

a. b. c. d. e. f.

1. wer (who) + + + + + +

2. wasu? (what) + + + + + +

3. wasOC (what) + + + + +

4. wasic (what) + + + + + +

5. was^ (what) + + + + + +

6. wo (where) + + + + + +

7. woher (whence) + + + + + +

8. wohin (whither) + + + + + +

9. wann (when) + + (wenn) ? -

10. seit wann (since when) + + + - ? -11. bis wann (until when) + + + - ? -

12. warum (why) + + + — — —

13. weswegen (why) + + + + + -14. wieso (why) + + + - - -

15. wozu (/or u>Aaf) + + + - + -

16. une^g^iÄou;) + + + + ? —

17· wie^^how) + + + + + —

15

I S . wieSPZC (how) + + +

19. was für (what kind of) + + +

20. welch (what) - + +

2 1 . welcher (which) + + +

22. wieviel (how much/many) + + +

Several interest ing things can be learned from this overview. F irs t , the no-matter-conditional and the exclamatory clause constructions seem to be very general, being the only ones i n which a l l 22 listed wh-words occur. Second, there seems to be a clear cut between the f i rst 8 words and the rest, since only these occur i n a l l six functions. Especially the weak indefinite is possible only as argument or i n local function, not i n temporal, causal, final, modal, or determiner function (its stressable irgencf-counterparts, however, do occur i n some of these functions: irgendwann, irgendwie, irgendwelche). The same holds tendentiai ly for the phrasal constructions (rows d and e); where they occur, they are mostly better w i t h a resumptive pronoun (Weswegen Max unzufrieden ist, (deswegenP.Q) kündigt er auch (Why Max is discontented, (therefore/1 Q) he gives notice)', Wie man sich bettet, (so/?0) liegt man (How one lies down, (so/10) one lies). T h i r d , our grouping of the six types into three groups seems to be just i f ied . Let us now have a look at the semantics of each construction i n t u r n .

3.3. Lexical wh-constructions

The weak indef ini te pronouns wer, was, wo, woher, wohin behave semantically exactly l ike the i r stressable counterparts jemand/irgendwer/irgendjemand, etwas/irgendetwas/irgendwas, irgendwo, irgendwoher, irgendwohin, i.e. they correspond, according to one's preferences to a (restricted) existent ial quant i f ier or to an un fami l iar i tem introducer. I indicate the semantics i n the former style and w i t h a Si tuat ion Semantics inspired notat ion, whose syntax mimics SOV-languages: first the arguments, then the predicate, followed by the polarity . Interpretations, i.e. t r u t h conditions, are relative to a given s i tuat ion s, so " i n s: o" is to be read as "the state of affairs σ holds i n s".

(46) Da singt wer.

(There is somebody singing.)

16

(46a) i n s: 3x[<xfperson',l > & <x,sing%\\ >]

M u l t i p l e weak indefinites are impossible since they cannot occur several times i n themselves, but weak indefinites can of course cooccur i n one clause and pose no special problem for the analysis (neglecting tense again):

(47) Da hat wer was gesagt.

(Somebody has said something.)

(47a) i n s : 3xy[<x,person*\l > & <y,person'\0> & <x,y,sag%\\>]

3.4. Phrasal wh-constructions

3.4.0 Overview

Phrasal wh-constructions consist of a wh-word followed by a clause where an expression of the same category (and case) as the wh-word is missing. The whole t h i n g has again the same category (but not necessarily case). E.g. i n Eva zerbrochen hat (Eva has broken), a non-personal NP i n the accusative case is missing, so was Eva zerbrochen hat (what Eva has broken) is the corresponding wh-phrase, and Max repariert, was Eva zerbrochen hat (Max is fixing what Eva has broken) as well as Was Eva zerbrochen hat, war nicht leicht zu reparieren (What Eva has broken was not easy to fix) are sentences containing that phrase i n an accusative and a nominative position, respectively.

Since was can be both accusative and nominative, the case m a r k i n g does not interfere w i t h the grammatical function of the wh-phrase i n the main clause. The s ituation is different w i t h the personal wer (nominative) vs. wen (accusative). Here often a left dislocation construction is used to avoid a bad feeling about the 'wrong' case m a r k i n g , cf. (48), where interna l and external case (case of the slot i n the embedded clause and case of the wh-phrase i n the matr ix clause) coincide, w i t h (49), where there is a difference:

(48) Wer Max berät, findet ihn sympatisch.

(Whoever counsels Max, likes him.)

(49) Wen Max berät, (der) findet ihn sympatisch.

(Whoever Max counsels, (he) likes him.)

W h a t is the meaning of a wh-phrase, what does i t denote? I c la im that i t denotes the max imal ent i ty which satisfies both the open sentence of the wh-phrase and the restriction imp l i c i t i n the wh-word. So, wen Eva kennt denotes the person(s) Eva knows and was Eva

17

kennt denotes the non-person(s) she knows, wo Paul wohnt denotes the location(s) of Paul's home and wohin dieses Flugzeug fliegt denotes the destination(s) of this airplane. As can be seen from the examples, wh-phrases do not care about semantic number: They are syntactically singularia tantum, but that does not say anyth ing about the cardinal i ty of what they denote.

So one desideratum for a formal representation is that i t be able to express this lack of specificity w i t h respect to the cardinal i ty of the denotatum. We could do this w i t h a sets-only domain, t reat ing individuals as singletons, b u t for reasons which are nicely explained in part 3.1 of L ink [5 ] , I prefer the lattice-theoretical approach, and I represent the denotation of wen Eva kennt (whom Eva knows) as Σχ[<χ,person'; 1 > & <e,x,kenn'\\>], i.e. the sum of individuals that are both persons and known by Eva.

Since wh-phrases can also function as declarative clauses, interrogative clauses, in f in i t i ve phrases, temporals and locatives/directionals, we have to extend Link's semi-lattices to these cases as wel l . Here, some interesting questions arise: I f we decide to let wh-phrases in clausal positions denote sums of states of affairs, where does the difference between declaratives and interrogatives go? And wouldn't the most natura l interpretation of a j o i n semi-lattice of states of affairs be their adjunction? But i f Max wants to know just whether i t is cold and whether i t is ra in ing , what Max wants to know doesn't denote the state of affairs that i t is cold or i t is ra ining . So there are several reasons, which I don't want to discuss here, supporting the assumption that the meanings of the daß and ob, which introduce the standard embedded declaratives and interrogatives, can be best represented by two related but different functions that take the state of affairs denoted by the clause as argument and bui ld a new object from which regular sums can be bu i l t . The functions are fact and issue, they Objectify' states of affairs and are characterized by the fo l lowing constraints and def init ion:

For any situations s, s \ and state of affairs o:

(A) i n s: <fact(o),sf,holds4n;l> i f f i n s \ ο

(Β) i n s: <non-facKo),s',holds4n\l> i f f i n 5* ορρ(σ), where ορρ(σ) is l ike σ except for the inverted polarity.

(C) issue(o) := fact(o) + non-fact(o)

(The plus-sign is used for the join-operator.)

We thus get the following representations for a declarative and an interrogative clause, respectively:

(50) Daß es regnet, steht fest.

18

(That it is raining is a fact.)

(50a) i n s: < facti < regri\ 1 >),feststeht 1 >

or, equivalent ly 1 :

(50b) i n s: <Vx[x = facK<regn\l >)], feststeht >

(Here, of course, the sum-formation is redundant and emphasizes only the paral le l ism w i t h the fol lowing case.)

(51) Ob es regnet, ist fraglich.

(Whether it is raining is debatable.)

(51a) i n s : <issue(<regn';l>),fraglich'\l>

This amounts, given the de f in i t ion (C), to (51b) or, equivalently , (51c):

(51b) i n s : <[fact{<regn%\\» + non-fact«regri\\>)\,fraglich'\\>

(51c) i n s : <Vx[[x = fact(<regn',l>)]\/[non'fact(<regnt\l>)}lfeststeht%l>

Issues are of course inconsistent, i.e. the conjunction of the ir parts cannot hold i n any real s i tuation, but that does not do any harm, on the contrary, i t seems to be exactly what is required i n the argument position of interrogative embedding predicates, since independent of the polarity of the relevant state of affairs, these predicates are always impl i c i t l y also about its opposite. More on interrogatives below.

To come back to our NR's, we can say now that they denote sums in general: Sums of persons i n the case of personal terms, sums of other objects i n the case of impersonal terms, sums of locations and times w i t h locatives and temporals, and sums of facts and issues w i t h declaratives and interrogatives.

M u l t i p l e lexical wh-constructions are not possible for syntactic reasons, but mult ip le phrasal wh-constructions are syntactical ly possible (cooccurrence of several wh-words re lat ing to or i n the same clause), they just don't occur, and here is something to be explained.

3.4.1. Naked relatives

W i t h these general remarks about the semantics of wh-phrases, the analysis of naked relatives is already completed, what follows are examples which show what is to be expected from the analysis ( ignoring tense):

(52) Wer da singt, ist musikalisch.

(Whoever is singing there is musical.)

19

(52a) i n s : <I,x[<x,persori\\> & <x,sing%,\ >],musikalisch*,l>

(53) Was Eva gekauft hat, ist teuer.

(What Eva bought is expensive.)

(53a) i n s : <Ex[<x ,person , ; 0> & <e,x,kauf\l>],teuer*,l>

(Note that i t does not follow that i f Eva bought, among other things, matches, matches are expensive, i f we decide to let expensive be a non-distr ibutive predicate , or even an antipersistent one 4 as we probably should.)

Since NRs are just special members of the same category as the slot i n their ( internal) clause, they are expected to cooccur i n the same (external) clause and i n one another, and so they do:

(54) Wer In'sein will, sagt, was alle sagen.

(Whoever wants to be \n'says what everybody says.)

(55) Wo wer tut, was er will, nicht zurechtgewiesen wird, herrscht Chaos.

(Whereever whoever does whatever he wants is not reproved, chaos

rules.)

3.4.2. Pseudo-cleft sentence subjects

According to Halvorsen ([3],p.D, a pseudo-cleft sentence is "a specificational copular construction w i t h a free, or headless, relative clause as subject". (He distinguishes specificational copular constructions such as The murderer is the gardener from predicational copular constructions such as The murderer is vegetarian.) I f this assumption is correct, the ambigui ty i n (56) is due solely to the predicate, and we need not worry about a special treatment of pseudo-cleft sentence subjects, since we can assign them exactly the same representation as a l l other NRs.

(56) Was Max gefallt, ist Unsinn.

(What Max likes is nonsense.)

I n the pseudo-cleft reading, (56) means that M a x likes nonsense, i n the other reading, i t means that Max likes something, and t h a t happens to be nonsense. The trouble is that the predicate noun cannot be the source of the ambiguity either, leaving the copula as its only possible remain ing source, and many l inguists are reluctant to accept this option. There are,

20

however, independent arguments for such an assumption (cf. Doron[2]), and we w i l l accept i t for our purposes, w r i t i n g < a , r ; l > for the state of affairs where a exemplifies r, and < a , r , e q ; l > (defined as < a , r , = ; l > ) for the state of affairs where a is identical w i t h r. So we get:

(57) Wer da singt, ist Ευα und Max.

(Who is singing is Ευα and Max.)

(57a) i n s : <Zx[<x,persori,\> & <x,sing'tl>],e + m,eq,l>

There is a problem w i t h number agreement i n German here, which I don't want to go into here, instead I want to point out how nicely our sum analysis of NRs fits i n w i t h conjoined predicate nominals.

3.5. Clausal wh-constructions

Since the difference between lexical and phrasal wh-constructions is obvious, and the assumed differences among the latter disappeared, the m a i n task for the analysis seems to account for the differences both between these two and the wh-clauses, and for those among the latter.

3.5.1. Interrogative wh-clauses

Embedded interrogative wh-clauses should be related adequately to o6-interrogatives for which an analysis has already been presented above (3.4.0), since they f i t into the same slots as these and are semantically related. 06-clauses have been analyzed as denoting issues, i.e. sums of the objects denoted by the corresponding cfa/i-clause and its negative counterpart:

(58) Eva weiß nicht, daß jemand singt.

(Eva does not know that someone is singing.)

(58a) i n s : <e,fact(3x[<x,person%\\> & <Jt , smg ' ; l>]) ,u ; iss* ;0>

(59) Eva weiß nicht, ob jemand singt.

(Eva does not know whether someone is singing.)

(59a) i n s : <e,issue(3x[<x,person',l> & <x,sing'\l>l),u;/ss*;0>

Since embedded wh-interrogatives can be ( in fact, should be, cf. [101) considered as generalized οό-interrogatives, i t is now almost obvious how to represent the ir meaning:

(60) Wer singt, weiß Eva nicht.

21

(Eva does not know who is singing.)

(60a) i n s: <e£y 3x[<xtperson',l >

& [y = issue(<xtsing'\l>)] ],wiss'',0>

A n analogous representation of mult ip le wh-interrogatives poses no problem:

(61) Wer wen wann womit bestochen hat, ist noch völlig offen.

(Who bribed whom when with what is still completely open.)

Independent interrogatives should be treated analogously, I w i l l not go into the details here and only mention that the relevant notion which should be provided by any account of their meaning, t h a t of a complete direct answer (cf. ZaeffererflO], p.73ff.), or most informative answer (cf. Peters[8), p .24l ) , can be easily defined as any maximal consistent conjunction of the states of affairs that are determined by the denotation of the sentence, i.e. one that holds i n a s i tuat ion s i f f s decides a l l and only the issues in that denotation, where 's decides an issue x* means One part of χ holds in s\

These conjunctions correspond U> the s i t u a t i o n s t h a t i n P e t e r account are i n the range of the re lat ion denoted by the interrogative. Informative answers are then those which state at least one member of such a conjunction, i.e. determine a s i tuat ion which decides at least one of the issues i n question. The presupposition of a question is the conjunction of a l l states of affairs that are entailed by a l l members of the given set of states of affairs.

This rules out the often assumed existential presupposition of wh-interrogatives, correctly, i n my view, since i t should be treated rather as an impl icature (cf.[10], p.78f.), which can be derived from the fact that i n general i t doesn't make much sense to ask a wh-question (no matter what i ts polarity) wi thout expecting at least one of the issues to be decided as factual .

Biased questions on the other hand are those where at least one of the issues (no matter what its polarity) is expected to be decided as non-factual, y ie ld ing the expected answers It is not raining for Is it (really) raining? and It is raining for Is it (really) not raining?

Rhetorical questions are on this account nothing else than strongly biased questions, i.e. questions where i t is implicated that a l l issues (no matter what the ir polarity) are decided as non-factual, y i e ld ing Nobody knows this book as impl icature of Who knows this book? and Everybody knows this book as implicature of Who doesn 9t know this book?

3.5.2. Exclamatory wh-clauses

22

The s i tuat ion seems to be quite s imi lar w i t h embedded wh-exclamatories, w i t h three major differences: F i rs t , as noted before, they are interchangeable w i t h da/?-clauses, but not w i t h 06-clauses. Second, the informational content is very weak, being not more than the fact that χ has P, where χ are those that have P. Th i rd , the embedding sentence does typical ly not entai l one w i t h a semantically weaker wh-exclamatory. Thus, (62) below, taken together w i t h the assumption that Max is among the singers, does not entai l (63), a lthough the paral le l inference w i t h interrogatives seems to hold, since i t could be that i t is the composition of the group of singers, which makes its s inging amazing, or the membership of somebody else than Max. (For a somewhat more detailed argumentation see Zae f ferer [ l l l . ) I t seems to me that these three properties have to be accounted for in two different places.

The f i rst two can be captured by assuming that wh-exclamatories are but generalized cfa/?-exclamatories jus t as wh-interrogatives are but generalized 06- interrogatives (as we did already i n 1.2 above) and that the generalization does not r u n over the total sum of facts that have the required property, but only those that in addition to that hold i n s. I represent t h a t by an indexed sum bu i ld ing operator Σ 5 , which is defined as follows: Σ. ;ν[Φ] : = Ey[ <y,s,holds 4n\\ > & Φ). This yields the following representation:

(62) Wer da singt, (das) ist erstaunlich.

(Who is singing is amazing.)

(The pronoun i n parenthesis serves to rule out the NR reading )

(62a) i n s: <Σ%γ 3x[<x,person%\\ > & [y = facti <x,sing',l >)]],erstaunlich%\\ >

(63) Daß Max da singt, ist erstaunlich.

(That Max is singing is amazing.)

(63a) i n s : <fact(<m,singi,\\>),erstaunlich%\l>

The second property can be captured by the assumption that wh-exclamatory embedding predicates (at least a l l I know of) are non-distr ibutive, jus t l ike expensive i n the example (53) above. So we get the inference from (62) to (63) jus t i n case Max is the only singer. This assumption is supported by the fact that the same non-d is tr ibut iv i ty can be observed when wh-exclamatory embedding predicates occur w i t h p l u r a l MPs: The achievements of the working group are amazing can be true even i f no single achievement of the group, taken by itself, is amazing.

The fo l lowing example shows that mult ip le wh-exclamatories are possible and pose no problem for the proposed analysis:

23

(64) Max wundert sich darüber, wer wo wohnt.

(Max is astonished at who lives where.)

(64a) i n s: < m , 3x,l[<x,person',l > & < I,location9,! > &

[y = fact(<x,l,wohnt,l>)]], sich-wunder-iiber%\\>

Independent exclamatories, except maybe for degree exclamatories such as 1st das aber leicht! (Boy, is that easy!), can be treated analogously i f one assumes that here a f i t t i n g att i tude expressing predicate (which is also non-distributive) takes the place of the superordinate clause.

3.5.3. Antecedents of no-matter-conditionals

These are the one the analysis of which I am most unsure about. They share a lot of properties w i t h both interrogatives and antecedents of regular conditionals, b u t differ from both of them i n interesting ways. Whereas interrogatives come i n three varieties, yes-no, alternative , and wh-interrogatives, (ob Eva kommt (whether Eva comes), ob Max oder Eva kommt (whether Max or Eva comes), wer kommt (who comes)), no-matter-conditionals lack the last option - (65) and (66) are o.k., (67) is not:

(65) Wer auch kommt, ich gehe.

(Whoever comes, I go.)

(66) Ob Max kommt oder nicht/Ob Max kommt oder Eva, ich gehe.

(Whether Max comes or not/Whether Max comes or Eva, I go.)

(67) *ObMax kommt, ich gehe.

(^Whether Max comes, I go.)

Note furthermore that the wh-variety seems to need a generalizing particle i n order to qual i fy as a no-matter-conditional antecedent:

(68) Wann du auch kommst/Wann immer du kommstHWann du kommst, du bist willkommen.

(No matter when you come/Whenever you comeP.When you come,you are welcome.)

24

The most notable difference from regular conditionals consists i n the word order of the consequent, ind icat ing t h a t i n the latter , the antecedent forms a single clause w i t h the consequent, whereas i n t h e i r no-matter-counterparts, the antecedents are only paratactically adjoined to the consequent (they are both syntactically and semantically dispensable, as i t were). So they behave syntactical ly exactly l ike the antecedents of i l l ocut ionary conditionals, cf. (69)/(70) for an i l l u s t r a t i o n or the i l locutionary/regular conditional d is t inct ion :

(69) Wenn du das Buch suchst, es liegt (*dann) dort drüben.

(If you are looking for the book, (*then) it is over there.)

(70) Wenn du das Buch suchst, (dann) liegt es dort drüben.

(If you are looking for the book, (then) it is over there.)

The difference between (69) and (70) seems to be most s tra ight forwardly accounted for by the assumption that (69) is a conditional assertion, whereas (70) asserts a conditional , i.e. (70) encodes only one i l locut ion, whereas (69) consists of two i l locutions, one of expressing a condition and one of asserting. F r o m this a n a l y s i s one m i g h t expect t h a t the consequent c a n

also be a nondeclarative sentence, and indeed, this is the case:

(71) Wenn du Tschechisch kannst, was heißt Tiovno*?

(If you know Czech, what does liovno 'mean?)

Here, i t doesn't make much sense to assume that the antecedent is under the scope of the question operator.

This suggests a n analysis also of no-matter-conditionals as expressing conditional assertions, but since the t r u t h of the (semantically weak) antecedent is generally taken for granted, i ts effect is an emphasized assertion instead of a relativized one. The antecedent has the in terna l structure of a n interrogative, but i t is not i n the scope of a question operator, i t is l i n k e d instead by the two-place cond operator to the assertion that is i ts consequent (to remain consistent, we wr i te the operator last, the antecedent next to i t , and the consequent first):

(72) Wer da auch singt, Eva hört nicht zu.

(Whoever is singing, Eva does not listen.)

(72a) i n s: < < <etzuhor9;0>,assert;! >,

Σ;ν 3x[<x,person',l> & [y = issue(<x,sing',l >)] ] , cond\\>

25

This represents the state of affairs where i t is asserted t h a t Eva does not listen, on the condition of a sum of issues about somebody singing. But what does i t mean for a state of affairs σ to be conditional on a sum of issues? I t means, I suggest, t h a t any argument of any issue i n that sum involves σ:

in s: < σ . Σγ[Φ\,€οηά;1> i f f

i n s: < a , t , inuolved-by;l> for any state of affairs t such that issued) is part of

Σγ[Φΐ

Thus, the i n t u i t i o n that (73) follows from (72), provided Max and Fr i t z are persons, is formally captured.

(73) Ob Max singt oder Fritz, Eva hört nicht zu.

(Whether Max is singing or Fritz, Eva does not listen.)

Unfortunately , this would also make the inference from (72a) to to the formal counterpart of (74) f ormal ly valid (provided Max is a person), wh i ch is of course inadequate.

(74) *Ob Max singt, Eva hört nicht zu.

(^Whether Max is singing, Eva does not listen.)

But i f we assume that a no-matter-conditional (by contrast to a regular one) requires an alternative , i.e. an at least two-part sum of issues as antecedent, we can rule this out and explain the fact t h a t the yes-no interrogatives do not have a no-matter-conditional antecedent counterpart. This assumption is supported by the fact that i n German, no-matter-conditional antecedents are expressable by pref ixing the predicate egal (equally) to a w h - or οό-clause, a predicate that is inherent ly relational , i.e. at least two-place. Hence (74) is out, since there is noth ing to which Max's s inging is compared, but (75) is fine:

(75) Ob Max singt oder nicht, Eva hört nicht zu.

(Whether Max is singing or not, Eva does not listen.)

This seems to be an at least plausible explanation for an otherwise total ly mysterious lack of paral lel ism. I w i l l conclude this section w i t h an example t h a t shows the possibility of mul t ip le wh-antecedents of no-matter-conditionals, and hence shows one more property shared w i t h interrogatives:

(80) Wer auch von wem womit bestochen wurde, ich glaube nicht mehr

an die Integrität unserer Regierung.

26

(Whoever has been bribed by whom with what, I don't believe anymore

in the integrity of our government.)

3.6. What relates naked relatives to their kin?

Looking back at our analysis of the semantics of German wh-constructions, we are able now to identify both the ties that hold the family together and the specific features t h a t characterize its members. I n al l six cases, there is a clause and at least one wh-word m a r k i n g at least one slot i n i t which is the position of a major constituent. The clause w i t h the slot(s) i n i t can be thought of as denoting a condition. The question is: W h a t does the wh-construction denote?

I n the f i rst case, that of the weak indefinite, i t denotes simply an indeterminate ind iv idual , and the whole clause denotes the state of affairs where the condition holds of an indeterminate ind iv idual .

I n the second case, that of the n a k e d r e l a t i v e s , it denotes a d e t e r m i n a t e I n d i v i d u a l ,

namely the ind iv idua l which is the sum of a l l the (indeterminate) individuals of which the condition holds.

The t h i r d case, that of the pseudo-cleft subjects, is in te rna l l y exactly l ike the second case, i t was suggested, and that means that a l l the differences that can be found are explainable i n terms of the special role they have, namely as arguments of equational sentences.

Case number four, the exclamatories, is different. Here, not the sum of those individuals is b u i l t that meet the condition, but the sum of factual instantiations of meeting the condition, i.e. of factual states of affairs, or, more precisely, the sum of fact objects corresponding to these states of affairs.

Cases five and six, interrogatives and wh-antecedents of no-matter-conditionals, are s imi lar to the exclamatories in that they denote sums of instantiat ions of meeting the condition, i.e. of states of affairs, but here, the states of affairs need not be factual and the sum contains issues, i.e. each fact object corresponding to an instant iat ion of meeting the condition comes paired w i t h its negative counterpart. Aga in , as w i t h cases two and three, the i n t e r n a l structure seems to be identical , so i t seems just i f ied to assume t h a t differences can be shown to be derivable from differences i n the role they play in the larger structures they are a part o f

A t the end of paragraph 3.4.0 i t was claimed that the contrast between phrasal and clausal wh-constructions w i t h respect to the possibli l ity of mul t ip le non-conjoined wh-words

27

occurring at the same level of embedding (absent i n both of the phrasal types, present in a l l three of the clausal types) is something which is to be explained. The fami ly picture just out l ined seems to provide a nice explanation:

According to this picture, phrasal wh-constructions denote the sum of individuals of which the condition holds. This can be a one-place condition only, since many-place conditions would determine several sums, which could not be denoted by a single constituent wi thout being integrated into a larger whole. That is why Wer wen liebt, der neckt sich (Whoever loves whom, teases himself) is out, whi le Was sich liebt, das neckt sich (Lovers tease each other) is i n .

Clausal wh-constructions, on the other hand, denote basically sums of states of affairs, i n th is picture, and here, i t does not matter whether these states of affairs come from instantiat ions into one-place or many-place conditions, y ie ld ing one definite sum i n any case.

Notes

* Univers i ty of Munich and Center for the Study of Language and Information, Ventura H a l l , Stanford, Cal i fornia 94305 USA.

Copyright (c) 1986 Dietmar Zaefferer

This paper was w r i t t e n under a grant from the DFG (Deutsche Forschungsgemeinschaft, German Research Association). I 'd l ike to thank the people at CSLI , especially John Perry, for their generous hospital i ty , Carol Kiparsky , Lynne Seymour, and D i k r a n Karagueuzian for technical assistance, and especially Michael Böttner, Godehard L i n k , and Stanley Peters for their most insp i r ing comments on earl ier versions. Whatever remains to be desired i n this paper is only my fau l t , not theirs.

1. For purposes of readabil i ty , I use in f ix notation wi thout explicit polar i ty and w i t h square brackets for identit ies that do not translate a surface copula.

2. I intent ional ly do not t a l k about in f in i t i ve phrase NRs here, since this would be a longer side isssue. 3. Dis tr ibut ive predicates ( in Link 's sense) are true of a l l atoms i n a sum, i f they are true of a sum. 4. Anti -persistent predicates ( in Peters* sense) are true of a l l parts (subsums) of a sum, i f they are true

of a sum.

References

[1] Carlson, Greg. A Uni f ied Analysis of the Engl ish Bare P lura l . Linguistics and Philosophy 1:413-457,1977.

28

[2] Doron, Edit . The Semantics of Predicate Nominals. Submitted to Language.

[31 Halvorsen, Per -Kr is t ian . The Syntax and Semantics of Cleft Constructions. Department of Linguistics, Univers i ty of Texas at Aust in , 1978.

[4] Lewis, David K. Adverbs of Quantif ication. I n Edward Keenan (editor), Formal Semantics of N a t u r a l Language, pages 3-15. Cambridge Univers i ty Press, 1975.

[5] L i n k , Godehard Plura l . I n A r n i m von Stechow, Dieter Wunder l i ch (editors), Handbuch Semantik. Athenaeum, forthcoming.

[6] Locke, John A n Essay Concerning H u m a n Understanding. Dover, 1894.

[7] Montague, Richard. The Proper Treatment of Quanti f icat ion i n Ordinary English. I n Richmond H . Thomason (editor), Formal Philosophy, chapter 8, pages 247-270. Yale Univers i ty Press.

[8] Peters, Stanley. The Si tuat ion i n Question. I n Robinson Schneider et a l . (editors), Parasession on Nondeclaratives, pages 237-243.

29

Chicago Linguistic Society, 1982.

[9] Savigny, Eike von. Bemerkungen zur Rolle der inneren Wahrnehmung i m englischen Empirismus. Archiv fuer Geschichte der Philosophie 55:36-46, 1973.

[10] Zaefferer, Dietmar. Frageausdrucke und Fragen i m Deutschen. Zu ihrer Syntax, Semantik und Pragmatik ( = Studien zur theoretischen L inguis t ik 2) F i n k , München, 1984.

[11] Zaefferer, Dietmar. The Semantics of Non-Declaratives: I n v e s t i g a t i n g G e r m a n Exclamatories.

I n R. Bäuerle, Ch. Schwarze, A. von Stechow (editors), Meaning, Use, and Interpretat ion of Language, pages 466-490.

de Gruyter , Ber l in , 1983.

C S L I Publications R e p o r t s

1. R e s e a r c h P r o g r a m on S i t u a t e d L a n g u a g e (out of print)

2. T h e S i t u a t i o n i n L o g i c - I Jon Barwise

3. C o o r d i n a t i o n a n d H o w to D i s t i n g u i s h C a t e g o r i e s Ivan Sag, Gerald Gazdar, Thomas Wasow, and

Steven Weisler 4. B e l i e f a n d I n c o m p l e t e n e s s

K u r t Konolige 5. E q u a l i t y , T y p e s , M o d u l e s a n d G e n e r i c s for L o g i c

P r o g r a m m i n g Joseph Goguen and Jose Meseguer

6. L e s s o n s f r o m B o l z a n o Johan van Bent hem

7. S e l f - p r o p a g a t i n g S e a r c h : A U n i f i e d T h e o r y of M e m o r y

Pentti Kanerva 8. R e f l e c t i o n a n d S e m a n t i c s in L I S P

Brian Cantwell Smith 9. T h e I m p l e m e n t a t i o n of P r o c e d u r a l l y R e f l e c t i v e

L a n g u a g e s J i m des Rivieres and Brian Cantwell Smith

10. P a r a m e t e r i z e d P r o g r a m m i n g Joseph Goguen

11. M o r p h o l o g i c a l C o n s t r a i n t s o n 3 c a u d \ u a v \ a a Tuitc A c c e n t

Meg Withgott and Per-Krist ian Halvorsen 12. P a r t i a l i t y a n d N o n m o n o t o n i c i t y i n C l a s s i c a l L o g i c

Johan van Benthem 13. S h i f t i n g S i t u a t i o n s a n d S h a k e n A t t i t u d e s

Jon Barwise and John Perry 1 4 - C . A s p e c t u a l C l a s s e s i n S i t u a t i o n S e m a n t i c s

Robin Cooper 15. C o m p l e t e n e s s of M a n y - S o r t e d E q u a t i o n a l L o g i c

Joseph Goguen and Jose Meseguer 17. M o v i n g t h e S e m a n t i c F u l c r u m

Terry Winograd 18. O n t h e M a t h e m a t i c a l P r o p e r t i e s of L i n g u i s t i c

T h e o r i e s C . Raymond Perrault

19. A S i m p l e a n d Ef f ic ient I m p l e m e n t a t i o n of H i g h e r -o r d e r F u n c t i o n s i n L I S P

Michael P. Georgeff and Stephen F .Bodnar 20. O n t h e A x i o m a t i z a t i o n of " i f - t h e n - e l s e "

Irene Guessarian and Jose Meseguer 21. T h e S i t u a t i o n i n L o g i c - I I : C o n d i t i o n a l s a n d

C o n d i t i o n a l I n f o r m a t i o n Jon Barwise

22. P r i n c i p l e s of O B J 2 Kokichi Futatsugi, Joseph A . Goguen,

Jean-Pierre Jouannaud, and Jose Meseguer 23. Q u e r y i n g L o g i c a l D a t a b a s e s

Moshe Vardi 24. C o m p u t a t i o n a l l y R e l e v a n t P r o p e r t i e s of N a t u r a l

L a n g u a g e s a n d T h e i r G r a m m a r Gerald Gazdar and Geoff Pul lum

25. A n I n t e r n a l S e m a n t i c s for M o d a l L o g i c : P r e l i m i n a r y R e p o r t

Ronald Fagin and Moshe Vardi

26. T h e S i t u a t i o n i n L o g i c - I l l : S i t u a t i o n s , S e t s a n d t h e A x i o m of F o u n d a t i o n

Jon Barwise 27. S e m a n t i c A u t o m a t a

Johan van Benthem 28. R e s t r i c t i v e a n d N o n - R e s t r i c t i v e Modi f icat ion

Peter Sells 29. E q u a t i o n s , S c h e m a t a a n d S i t u a t i o n s : A

F r a m e w o r k for L i n g u i s t i c S e m a n t i c s Jens E r i k Fenstad, Per-Krist ian Halvorsen, Tore

Langholm, and Johan van Benthem (out of print) 30. I n s t i t u t i o n s : A b s t r a c t M o d e l T h e o r y for

C o m p u t e r S c i e n c e J . A . Goguen and R. M. Burstal l

31 . A F o r m a l T h e o r y of K n o w l e d g e a n d A c t i o n Robert C . Moore

32. F i n i t e S t a t e M o r p h o l o g y : A R e v i e w of K o s k e n n i e m i (1983)

Gerald Gazdar 33 . T h e R o l e of L o g i c i n A r t i f i c i a l I n t e l l i g e n c e

Robert C . Moore 34. A p p l i c a b i l i t y of I n d e x e d G r a m m a r s to N a t u r a l

L a n g u a g e s Gerald Gazdar

35 . C o m m o n s e n s e S u m m e r : F i n a l R e p o r t Jerry R. Hobbs, et al .

36. L i m i t s of C o r r e c t n e s s in C o m p u t e r s Brian Cantwell Smith

37. Ott tVkt: CoVittcncc bnd S t r u c t u r e of I>l*co\*r««>

Jerry R . Hobbs 38. T h e C o h e r e n c e of I n c o h e r e n t D i s c o u r s e

Jerry R. Hobbs and Michael H. Agar 39. T h e S t r u c t u r e s of D i s c o u r s e S t r u c t u r e

Barbara Grosz and Candace L . Sidner 40 . A C o m p l e t e , T y p e - f r e e " S e c o n d - o r d e r " L o g i c a n d

I t s P h i l o s o p h i c a l F o u n d a t i o n s Christopher Menzel

41 . P o s s i b l e - w o r l d S e m a n t i c s for A u t o e p i s t e m i c Logic Robert C . Moore

42 . D e d u c t i o n w i t h M a n y - S o r t e d R e w r i t e Jose Meseguer and Joseph A. Goguen

4 3 . O n S o m e F o r m a l P r o p e r t i e s of M e t a r u l e s Hans Uszkoreit and Stanley Peters

44. L a n g u a g e , M i n d , a n d I n f o r m a t i o n John Perry

46 . C o n s t r a i n t s o n O r d e r Hans Uszkoreit

47 . L i n e a r P r e c e d e n c e i n D i s c o n t i n u o u s C o n s t i t u e n t s : C o m p l e x F r o n t i n g i n G e r m a n

Hans Uszkoreit 48 . A C o m p i l a t i o n of P a p e r s on U n i f i c a t i o n - B a s e d

G r a m m a r F o r m a l i s m s , P a r t s I a n d I I Stuart M. Shieber, Fernando C . N . Pereira, Lauri

Karttunen, and Martin K a y 49 . A n A l g o r i t h m for G e n e r a t i n g Q u a n t i f i e r Scopings

Jerry R . Hobbs and Stuart M. Shieber 50. V e r b s of C h a n g e , C a u s a t i o n , a n d T i m e

Dorit Abusch 51. N o u n - P h r a s e I n t e r p r e t a t i o n

Mats Rooth 52 . N o u n P h r a s e s , G e n e r a l i z e d Q u a n t i f i e r s

a n d A n a p h o r a Jon Barwise

53 . C i r c u m s t a n t i a l A t t i t u d e s a n d B e n e v o l e n t 80. C o g n i t i o n

John Perry 54. A S t u d y i n t h e F o u n d a t i o n s of P r o g r a m m i n g 61 .

M e t h o d o l o g y : S p e c i f i c a t i o n s , I n s t i t u t i o n s , C h a r t e r s a n d P a r c h m e n t s 82 .

Joseph A . Goguen and FL M. Burstall 55 . Q u a n t i f i e r s i n F o r m a l a n d N a t u r a l L a n g u a g e s

Dag Westerstahl 56 . I n t e n t i o n a l i t y , I n f o r m a t i o n , a n d M a t t e r 83 .

Ivan Blair 57 . G r a p h s a n d G r a m m a r s

Wil l iam Marsh 84. 58 . C o m p u t e r A i d s for C o m p a r a t i v e D i c t i o n a r i e s

Mark Johnson 85 . 59 . T h e R e l e v a n c e of C o m p u t a t i o n a l L i n g u i s t i c s

Lauri Karttunen 86. 60 . G r a m m a t i c a l H i e r a r c h y a n d L i n e a r P r e c e d e n c e

Ivan A . Sag 87. 61 . D - P A T R : A D e v e l o p m e n t E n v i r o n m e n t for U n i f i ­

c a t i o n - B a s e d G r a m m a r s Lauri Karttunen 88.

62 . A S h e a f - T h e o r e t i c M o d e l of C o n c u r r e n c y Luis F . Monteiro and Fernando C . N. Pereira 89.

63 . D i s c o u r s e , A n a p h o r a a n d P a r s i n g Mark Johnson and Ε wan Klein

64. T a r s k i on T r u t h a n d L o g i c a l C o n s e q u e n c e 90. John Etchemendy

65 . T h e L F G T r e a t m e n t of D i s c o n t i n u i t y a n d t h e D o u b l e I n f i n i t i v e C o n s t r u c t i o n i n D u t c h 91.

Mark Johnson 66 . C a t e g o r i a l U n i f i c a t i o n G r a m m a r s 92.

Hans Uszkoreit 67 . G e n e r a l i z e d Q u a n t i f i e r s a n d P l u r a l s

Godehard Link 68 . R a d i c a l L e x i c a l i s m 93.

L a u r i Karttunen 69 . W h a t is I n t e n t i o n ?

Michael E . Bratman (See Bratman's Intentions, 94. Plans, and Practical Reason, Cambridge, Mass.: Har­vard University Press. 1987.) (out oj print) 95.

70. U n d e r s t a n d i n g C o m p u t e r s a n d C o g n i t i o n : F o u r R e v i e w s a n d a R e s p o n s e 96.

Mark Stefik, Editor 71. T h e C o r r e s p o n d e n c e C o n t i n u u m

B r i a n Cantwell Smith 97. 72 . T h e R o l e of P r o p o s i t i o n a l O b j e c t s of B e l i e f i n

A c t i o n 98, David J . Israel

73 . F r o m W o r l d s to S i t u a t i o n s John Perry 99,

74 . T w o R e p l i e s Jon Barwise 100,

75 . S e m a n t i c s of C l o c k s B r i a n Cantwell Smith 101,

76 . V a r i e t i e s of S e l f - R e f e r e n c e B r i a n Cantwell Smith (Forthcoming)

77 . T h e P a r t s of P e r c e p t i o n 102, Alexander Pentland

78. T o p i c , P r o n o u n , a n d A g r e e m e n t i n C h i c h e w a Joan Bresnan and S. A . Mchombo

79 . H P S G : A n I n f o r m a l S y n o p s i s 103. C a r l Pollard and Ivan A. Sag

T h e S i t u a t e d P r o c e s s i n g of S i t u a t e d L a n g u a g e

Susan Stucky (Forthcoming) M u i r : A T o o l for L a n g u a g e D e s i g n

Terry Winograd F i n a l A l g e b r a s , C o s e m i c o m p u t a b l e A l g e b r a s , a n d D e g r e e s of U η s o l v a b i l i t y

Lawrence S . Moss, Jose Meseguer, and Joseph A . Goguen

T h e S y n t h e s i s o f D i g i t a l M a c h i n e s w i t h P r o v a b l e E p i s t e m i c P r o p e r t i e s

Stanley J . Rosenschein and Leslie Pack Kaelbling F o r m a l T h e o r i e s of K n o w l e d g e i n A I a n d R o b o t i c s

Stanley J . Rosenschein A n A r c h i t e c t u r e for I n t e l l i g e n t R e a c t i v e S y s t e m s

Leslie Pack Kaelbling O r d e r - S o r t e d U n i f i c a t i o n

Jose Meseguer, Joseph A . Goguen, and Gert Smolka M o d u l a r A l g e b r a i c S p e c i f i c a t i o n of S o m e B a s i c G e o m e t r i c a l C o n s t r u c t i o n s

Joseph A . Goguen P e r s i s t e n c e , I n t e n t i o n a n d C o m m i t m e n t

Phi l Cohen and Hector Levesque R a t i o n a l I n t e r a c t i o n a s t h e B a s i s for C o m m u n i ­c a t i o n

Phi l Cohen and Hector Levesque (Forthcoming) A n A p p l i c a t i o n of D e f a u l t L o g i c to S p e e c h A c t

C . Raymond Perrault M o d e l s a n d E q u a l i t y for L o g i c a l P r o g r a m m i n g

Joseph A . Goguen and Jose Meseguer O r d e r - S o r t e d A l g e b r a S o l v e s t h e C o n s t r u c t o r -S e l e c t o r , M u l i t p l e R e p r e s e n t a t i o n a n d C o e r c i o n P r o b l e m s

Joseph A . Goguen and Jose Meseguer E x t e n s i o n s a n d F o u n d a t i o n s for O b j e c t - O r i e n t e d P r o g r a m m i n g

Joseph A . Goguen and Jose Meseguer L 3 R e f e r e n c e M a n u a l : V e r s i o n 2.19

William Poser C h a n g e , P r o c e s s a n d E v e n t s

Carol E . Cleland (Forthcoming) O n e , N o n e , a H u n d r e d T h o u s a n d Spec i f icat ion L a n g u a g e s

Joseph A . Goguen C o n s t i t u e n t C o o r d i n a t i o n i n H P S G

Derek Providian and David Goddeau A L a n g u a g e / A c t i o n P e r s p e c t i v e o n t h e D e s i g n of C o o p e r a t i v e W o r k

Terry Winograd I m p l i c a t u r e a n d D e f i n i t e R e f e r e n c e

Jerry R. Hobbs T h i n k i n g M a c h i n e s : C a n T h e r e b e ? A r e we?

Terry Winograd S i t u a t i o n S e m a n t i c s a n d S e m a n t i c I n t e r p r e t a t i o n i n C o n s t r a i n t - b a s e d G r a m m a r s

Per-Krist ian Halvorsen C a t e g o r y S t r u c t u r e s

Gerald G a z d a r , Geoffrey K . Pul lum, Robert Car­penter, E w a n K l e i n , Thomas E . Hukari, Robert D. Levine

C o g n i t i v e T h e o r i e s o f E m o t i o n Ronald A l a n Nash

104. T o w a r d a n A r c h i t e c t u r e for R e s o u r c e - b o u n d e d A g e n t s

Martha E . Pollack, David J . Israel , and Michael E . Bratman

105. O n t h e R e l a t i o n B e t w e e n D e f a u l t a n d A u t o e p i s -t e m i c L o g i c

Kurt Konolige 106. T h r e e R e s p o n s e s to S i t u a t i o n T h e o r y

Terry Winograd 107. S u b j e c t s a n d C o m p l e m e n t s i n H P S G

Robert Borsley 108. Too ls for M o r p h o l o g i c a l A n a l y s i s

Mary Dalrymple, Ronald M . K a p l a n , Lauri K a r t ­tunen, K i m m o Koskenniemi, Sami Shaio, Michael Wescoat

109. C o g n i t i v e S i g n i f i c a n c e a n d N e w T h e o r i e s of R e f e r e n c e

John Perry 110. A l e p h : A S y s t e m S p e c i f i c a t i o n L a n g u a g e

Terry Winograd 111. F o u r t h Y e a r R e p o r t o f t h e S i t u a t e d L a n g u a g e

R e s e a r c h P r o g r a m 112. B a r e P l u r a l s , N a k e d R e l a t i v e s , a n d T h e i r K i n

Dietmar Zaefferer 113. E v e n t s a n d " L o g i c a l F o r m "

Stephen Neale 114. B a c k w a r d s A n a p h o r a a n d D i s c o u r s e S t r u c t u r e :

S o m e C o n s i d e r a t i o n s Peter Sells

115. T o w a r d s a L i n k i n g T h e o r y of R e l a t i o n C h a n g i n g R u l e s in L F G

Lori L e v i n (Forthcoming)

11 . Working Papers in Grammatical Theory and Discourse Structure: Interactions of Morphology, Syntax, and Dis­course

M . Dda, S . Wechsler, and D . Zee ( E d s . ) with an Introduction by Joan Bresnan

12 . Natural Language Processing in the 1980s: A Bibliography

Gerald Gazdar, Alex Franz, Karen Osborne, and Roger Evans (Forthcoming)

13 . An Information-Based Syntax and Semantics C a r l Pollard and Ivan Sag (Forthcoming)

14. Non-Well-Founded Sets Peter Aczel (Forthcoming)

15 . Outlines of Probabilism Richard Jeffrey (Forthcoming)

L e c t u r e N o t e s S e r i e s

The titles in this series are distributed by the Uni­versity of Chicago Press and may be purchased in academic or university bookstores or ordered directly from the distributor at 5801 Ellis Avenue, Chicago, Illinois 60637.

1. A Manual of Intensional Logic Johan van Benthem

2. Emotions and Focus Helen Fay Nissenbaum

3. Lectures on Contemporary Syntactic Theories Peter Sells Paper

4. An Introduction to Unification-Based Approaches to Grammar

Stuart M . Shieber 5. The Semantics of Destructive Lisp

Ian A . Mason 6. An Essay on Facts

Ken Olson 7. Logics of Time and Computation

Robert Goldblatt 8. Word Order and Constituent Structure in German

Hans Uszkoreit 9. Color and Color Perception:

A Study in Anthropocentric Realism David Rüssel Hilbert (Forthcoming)

10. Prolog and Natural-Language Analysis Fernando C . N . Pereira and Stuart M. Shieber