Theme, Focus and Negation. CES 5
 　（千葉大学ユーラシア言語文化論集５，2002）CES 5

Theme, Focus and Negation1)                                         　　　　　　　 　　1. Theme in Japanese
 Jens Rickmeyer (Rickmeyer 1999) criticized me for the interpretation of Japanese postposition wa:
“Pragmatisch interpretiert, reflektiert die Funktion f in … [f(x,p)] den verbalen Prozeß des Sprechers, daß er aus dem Gegenstandsbereich des gegenwärtigen Diskurses (des Elements) x (, das der mit wa formulierten Konstituente zugrundeliegt,) ausgewählt und über x ausgesagt, daß es um (die Proposition) p geht.” (Kaneko 1987)　(in the parentheses are Rickmyer’s supplements).
　Surely, in my interpretation at that time lack many important things; it said practically nothing more than the Chomskian functional definition of subject: subject is “NP in [ __, S]”  in the old era of the Standard Theory. The semantic interpretation of f in [f(x,p)] denotes the relation between x and p which reflects the traditional view that theme x is something the proposition p talks about. And the semantic restriction on x was far from being obvious, too. Namely, my assumption at that time is formulated roughly as

(1) theme is X in: f(__, TP),  
   where TP is “topic phrase” in the recent generative theory, and f is a function over X and TP
   in the meaning “TP talks about X”

And Rickmeyers comments concern the following three points:

(2)a. My formulation (1) lacks the verbal procedure of the the speaker who selects theme-
    phrase X out of the objects-class (Gegenstandsbereich) X*. He formulated it as
    ELEC(Dic, in X, ex X*).
  b. The objects-class X* belongs to the knowledge of the discourse partner in the present
    discourse situation. This situation is formulated as
    X*│SITU(...,__,...), and
  c. The proposition TP in (1) is related to X*, i.e. there has to be a reasonable congitive relation
    between the theme and the proposition about it.
    KOND((REID(X1,X│ELEC(Dic, in __, ex X*)),LOQI(Dic, Aud,Text:》vnX1...Yn《 ),    
    i.e. the condition has be guranteed for referencial reasonableness concerning the
    selection of X and the asserted propositon).

 I wrote somewhere that thematizable terms have to be stored in the mutual knowledge of the discourse partners, so that deixes like this one and temporal, local and personal terms like yesterday, at the corner and my wife can be the themes of ordinary sentences most suitably. On the contrary, a nominal phrase like a desktop PC is not suitable for the theme of a discourse in a high class ristorante when the partner are busy for choosing a good Toscana wine, but the same NP is often thematizable in a PC shop when we want to buy a notebook PC or a bigger one.
 From the very beginning of the discussion on the theme (cf. Kuno 1973, Kaneko 1975), it has been often remarked that a thematizable NP must be spezialized with the feature [+mentioned], i.e. it has to  be explicitely mentioned in the discourse beforehand. Difficulties arise when deixes and temporal terms can be easily thematized without being mentioned previously. Someone proposed to introduce a new feature [+known] to overcome the difficulties because this feature covers not only [+mentioned] terms, but also deixes in general and the generally known words which the speaker assumes as known to the hearer in the situation of present discourse. Of course, it cannot be applied to any known term in one’s individual encylopedia in the brain because a thematizable term is only what is relevant in the present discourse situation. It has to be stored in the present memory of the discours partners from the view point of the speaker. If this is not the case, the speech is sanctioned. Therefore, it is suitable to define  this pragmatic feature as [+stored in the present memory of the discours partners in view of the speaker’s assumption], in short, [+m].
But the feature [+m] alone is not sufficient to explain the usage of wa. Rickmeyer is right in talking about the selecion of X out of the objects class X*. But he forgets mentioning why X is selected from X* and what grammatical function X has in the context in which it occurs. He wants to explain his selection function with the following example:

(3) kyoositu=    de=wa        tabako  o   suw.ana.i=de kudasai.
   the class room in  theme     cigarette OBJ smoke-not   beg
   where his “=” shows a phrasal and “.” a morphematic concatenation.

 In the sentences (3), it is not obvious whether the wa-phrase is the theme of the sentence or a contrastive phrase. The latter NP-wa occurs when it is compared implicitely with other places, e.g. on an outdoor bench. Rickmeyer thinks that it is the case here. Namely, it is contrastively selected X out of X*, where X* the class of places where smoking is possible. It is not merely a NP-wa with the fearure [+m], but something more than it. But I do not agree with him, I rather assume that this NP-wa permits two ways of interpretations, namely it is ambiguous to be a thematic or a contrastive NP-wa as far as no further contextual informations are delivered. In order to approach the matter closer, let us examine the sentence (4), which contains many NP-wa’:

(4) kare wa  kyoo wa   gakusyoku       de  wa  boku to wa hanbaagu-lunch wa
　   he      today wa  Student-restaurent LOK wa  I  with wa hamburg-lunch wa  

　　 tabe-na-katta
　　　eat-NEG- PAST

 The first NP-wa is surely the theme of the sentence (4), but the following others seem to be contrastive; they are contrasted e.g. with yesterday, in a city restaurent with someone others and curry rice respectively. With the first NP-wa, the speaker intends to put his sentence into the chain of discourse remarking that he wants to talk about it. The thematic NP-wa, therefore, functions as a bridge between the previous and the present sentence which begins with the theme connecting them. It functions to determine the inter-sentencial constellation of the sentence, indicating an informational channel where the discourse should turns toward. The other obviously contrastive NP-wa’s in the sentence (4) have no intersentencial function, but have an intra-sententional function to indicate an informational weight of the terms. They often receive a heavy stress. A contrastive NP-wa denotes, namely a stressed term especially selected out of an object class.
 If we assume that a NP-wa in the leftmost position2) is thematic, but in other positions contrastive, the asumption contradicts not only Rickmeyer’s judgement that the NP-wa in (3) is contrastive, but also my idea about the ambiguity of its interpretation. The only way to avoid the contradiction is to assume that the thematic/contrastive interpretation depends on its position as well as on contextual information.
 Rickmeyer says about the contrastive usage of a NP-wa as follows: ”daß der kontrastive Gebrauch der Partikel als Sonderfall des oben diskutierten Gebrauchs erklärt werden kann, bei dem die Determinierung des Gegenstandsbereichs zu einem Teil durch den Postkontext ... expliziert wird”, i.e. the contrastive usage of a NP-wa is a special case of the selection of X out of X*, where the determination of X* is partially explicated by the “post-context”. He is surely right in saying that the selection of X out of X* is the fundamental procedure of thematization and contrastivization as well. But from syntactic view point, it is obvious that the leftmost order is the necessary condition for the thematic interpretation of a NP-wa. Of course, the contextual information has to be taken into account, too. I will not ask here what he means with “postcontextual explication” I think rather that a precontexual presupposition is more crucial for our problem.
Let us examine again the clear case of contrastive NP-wa’s in (4). As mentioned above, the second and the following NP-wa’s allow no thematic, but only contrastive interpretation. They are also regarded as terms which appear in the answer to the question (5):

(5) kare (wa) [1 kyoo] [2 gakusyoku de] [3 kimi to] [4 hanbaagu-lunch o] tabe-ta no?

The terms [1 ...] [2 ...] and [4 ...] are merely repeated in the sentence (4). But [3 kimi to] results from is the inevitable alteration according to the partner change in a discourse. This implies that the contrastive NP-wa’s are the directly prementioned terms Xi out of X* or an alternative selection out of the directly prementioned same object class X*. Let us put the conditions together into one feature [+direct premention], in short, [+dp]. We assume that the contextual feature [+dp] is the fundamental condition for the occurrence of a contrastive NP-wa. We can also establish the logical relation between [+dp] for a contrastive and [+m] for a thamatic NP-wa as follows:

(6) a. The thematic interpretation of NP-wa requires the condition:Xi [+m] in [__│X*].
　   b. The contrastive interpretation of NP-wa requires the condition Xi [+dp] in
　　　　[__│X*].
　   c. Xi [+dp]⊃Xi [+m], but not vice versa.

 Now, let us go back to Rickmeyer’s example (3). He maintains that the leftmost NP-wa in (3) is contrastive, but I think that it is ambiguous; it can be thematic or contrastive. This discrepancy will disappear, however, if we combine the syntactic condition with the semantic in the following order:

(7) a. NP-wa stands in the leftmost ⇒ thematic
   b. NP-wa denotes a familiar object to discourse partners, i.e.[+m] ⇒ thematic
   c. NP-wa is directly prementioned or it is an alternative selection, i.e.[+dp] ⇒ contrastive  

　To use Rickmeyer’s terms, the speaker chooses Xi out of X* and represents it as NP-wa in a sentence. It is interpreted as thematic in a contexually unmarked case if it goes through the conditions (7a) and (7b). But in a marked case in (7c), when other members of X*, e.g. in a smoking corner, or on an outdoor bench, or somewhere are the matter of particular concern for the speech partners, it is interpreted as contrastive. 

 Summing up the assumption so far, a NP-wa functions as the theme of a sentence or as the contrastive term. Both functions of NP-wa are ascribed to a locutional act to choose X out of X* as Rickmeyer formulated. For the successful performance of this act, there are the following conditions in morphosyntactic, lexical and contextual/situational fields:

(8) a. both thematic and contrastive usage of NP-wa originate from a locutional act:
     the selection of X out of [__│X*]
   b. lexical condition:NP has the semantic feature [±m]
     [+m] ⇒ thematic or contrastive
     [-m] ⇒ neither thematic nor contrastive
   c. morphosyntactic condition:NP-wa [±leftmost]
     [+leftmost] ⇒ thematic
     [- leftmost] ⇒ contrastive
   d. contextual condition:NP is in the context [+dp]
     [+dp] ⇒ contrastive

2. Theme versus Focus
 By means of the interpretative features of (8) we can assigned to the sentence (4) roughly the following description:

(4’) [TP[+m kare wa] [IP[+dp kyoo wa gakusyoku de wa boku to wa hanbagu-lunch wa] tabe-na-katta]]

   Now, let us exchange the negative predicate of the sentence (4) with an affirmative, we get a curious sentence like (9).

(9) [TP[+m kare wa] [IP[+dp kyoo wa gakusyoku de wa boku to wa hanbagu-lunch wa] tabe-ta]].

  The curiosity is due to the implications of the sentence: (i) the negative sentence (4) or the question (5) precedes the sentence (9) or (ii) this is followed by another negative narration for a revision like:

(10) ..., keredomo sonna hanashi wa    si-na-katta.
       but      such  a tale [+dp]  do-NEG-PAST

 This negative sentence is the necessary supplement to the affirmative (9), which is as a whole contrasted with it. On the contrary, the sentence with a negative predicate (4) needs not any comment like this. It seems to be self-satisfied. This implies that a contrastive NP-wa expects a negative predicate or a negative explanation in the post-context. A contrastive NP-wa seems to be congeninal with a negative circumstance because of its function to indicate the scope of negation as we see below.
 Explaining the theme/contrast-opposition and its pragmatic conditions, Rickmeyer says: “Zu guter Letzt bleibt noch die Frage, welches griffiges Etikett auf die Partikel=wa als repräsentativer Name für die Funktion geklebt werden kann, ohne daß es zu Divergenzen zwischen Deklaration and Inhalt kommt, wie es z.B. bei Thema oder Fokus der Fall war. Der Ausdruck durch Vorauswahl bedingter Geltungsbereich der Äußerung müßte also nur in eine latinisierter Kurzformnote10 gebracht werden.” In the note 10 he proposes as the best latin label for the function of wa, :”Vorgeschlagen ist bereits der Terminus Selektiv.” He defines the particle wa, namely, as the marker for theme or focus selected intentionally in the present situation. To the function of wa he proposes the label selective3). He is surely right in proposing the selection function ELEC in (2) above. But he overlooks the contextual conditions for the contrastive usage of NP-wa [+dp] above. Therefore, he does not distinguish theme from focus because he finds in them merely the same function of selection without mentioning its motivation.
 Coming back to the example (4), we find between the contrastive NP-wa’s and the negative predicate a special relation: what is negated in this sentence is not the thematic NP-wa, but only the contrastive ones in [+dp...]. In other words, the scope of the negation involves only the contrasted range, but excludes the thematic NP-wa.
 As for negation, negative elements, constituent (partial) negation, the scope of negation and especially the negation of quantification, etc., we have a long history of discussion, most of which have not yet been settled up. Although our problem is related to many of them, we will confine ourselves only to the necessary topics on the typology of contrast and focus.
 Let us examine again the negation of the thematic/contrastive ambiguity of the sentence (3) whose structure looks like as follows:

(11) [TP[+m/+dp? kyoositu de wa] tabako o [suw-[NEG ana-i]] de kudasai]].

If kyoositu de wa is a theme, the scope of the negative element ana-i includes the empty trace of the moved-out constituent like (12a). It is out of the scope of nagation. On the contrary, if the NP-wa is contrasted with some other locative terms precontexually or prosodically, it remains within the scope of nagation like (12b).

(12) a. [TP[TOP[location kyoositu de wa]][IP [location t][object tabako o] [V suw-] [NEG anai]]
    b. [IP[NEGP[location kyoositu de wa] [object tabako o] [[V suw-][NEG anai]]]

The structure (12b) is so clusmy that it cannnot be construed as a grammatical structure description. However, it coincides with an “internal logic” of the contrastive interpretation in the meaning that “it is not in the class room to smoke”. In a language without nominal negation like Japanese, a predicate negation underlies all sorts of negative structures and it appears in the surface within a verbal complex. This corresponds with the general idea that negation is a predicate, but not a quantification.
 Another argument says that the elements in the scope of negation can be regarded as the answer to the previous question, e.g.

(13) a. Q: kyoositu de tabako o sut-te mo ii desu ka?
    b.A1:[NEGP [VP [kyoositu de wa] [tabako o/wa] suw]-ana-i de] kudasai.
    c.A2:[VP kyoositu de tabako o sut-te wa] ikemasen.

In both sentences (13b) and (13c), the scope of negation is marked with wa. The sentence (13b) is a reqeust for the content of [NEGP...]. The verbal complex with negative adjective -ana-i involves both NP-wa’s in the sentence. The answer (13c) is a prohibition in the meaning:[IP[VP smoking in the class room wa] forbidden]. The VP as a whole is contained in the higher predicate of prohibition, where the particle wa marks the scope of the negation, namely, the [VP...]. Therefore, all wa’s in (13) mark the negated focus of the sentences both NPs in (13b) and the whole VP in (13c).
 Now, let us take a clear case of contrastive NP-wa’s. The following sentence includes three NP-wa’s:

(14) [TP [+m kare wa] [VP1 [+dp tabako wa] yame-ta] ga], [VP2 [+dp sake wa] kessite yame-na-i]  
    (He stopped smoking, but will never stop drinking.)

Here, [+m kare wa] is the thematic NP-wa for the whole sentence, but the other two are contrastive. It may be probable that a suprasegmental specification falls on the latter NP-wa’s, but even without such prosodic means, two NP-wa’s in [+dp ...] are obviously contrastive to each other. It is also the case, if both of them appear independent sentences as follows:

(15) [+m kare wa] [VP1 [+dp tabako wa] yame-ta] yo]. Keredomo [VP2 [+dp sake wa] kessite  yame-na-i].    

In both (14) and (15), contrastive NP-wa’s, [+dp tabako wa] and [+dp sake wa], are regarded as the focus of the VPs.  We can now summarize our assumption as follows:

(16) a. A thematic NP-wa [+m ...] is out of the scope of negation.
    b. A contrastive NP-wa [+dp ...] involved in the scope of negation is the focus of the sentence. 
    c. A contrastive NP-wa [+dp ...] functions as the focus in an affirmative sentence,
    d. A contrastive NP-wa [+dp ...] receives often a suprasegmental marking.

Therefore, we can assume the following:

(17) a. A NP-wa is ambiguous:it is thematic [+m ...] or contrastive [+dp ...].
    b. A contrastive NP-wa with the feature [+dp ...] functions as the focus both in a negative and in an
      affirmative sentence.

3. Morphological Focus-marking
 There is a language famous for a complex system of morphological focus marking. It is one of the “paleoasiatic” languages Yukagir, the structure of which has been recently investigated intensively in various papers of Endo Fubito and Nagasaki Iku. According to Nagasaki, it has three morphematic markings for focus in affirmative sentences of indicative mood: for the intransitive subject and the transitive object suffixes -leN, -ek 4) and for the predicate prefix5) me(r)-. The focussing is related closely to inflection paradigms. Yukagir has, according to Nagasaki (1997), at least eight different inflection paradigms: affirmative sentences in indicative mood alone are differentiated by five inflection types corresponding to what is focussed.

(18) a. if the focus falls on the intransitive predicate, this richest type of inflection paradigm is attached
      to the verb stem. Let us call this basic type as infl(I),        
    b. if the focus falls on the subject of the intransitive predicate, a simple inflection appears, namely
      infl(II),   
    c. if the transitive predicate is focussed, it appears infl(III) which is simple, too, but similar to
      infl(I),
    d. if the transitive object is focussed, a characteristic type of inflection appears, infl(IV), which
      resembles to infl(I), and    
    e. if the focus falls on the subject of the transitive predicate, the weakest type of
      inflection infl(V) appears, in which aorist-inflection isφexcept for 3pl.

Nagasaki shows further three types of paradigms for negative, imperative and interrogative moods. The negative inflection paradigm is almost identical to infl(I) except for 3sg/pl which is -φ/Nu respectively. We call it as infl(NEG).
Let us observe some typical examples cited from Nagasaki (1994 and others).

(19) a. focus on the intransitive subject
      ex.: ile-leN           qaldej-l.
         deer-focus   escape-infl(II)3sg
         (it is the deer who escaped)
    b. focus on the transitive object
      ex.: tudel ile-leN    pun’-mele.  
         3sg.  deer-focus  kill-infl(IV)3sg
         (it is a deer that he killed)
    c. focus on the intransitive predicate
      ex.: met mer-u:-jeN.
　         I   focus-go-infl(I)1sg
        (went I did)
    d. focus on the transitive predicate
      ex. met tet-ul         me-pot’esej-N.
         I   you-obj.case   focus-see off-infl(III)
        (I did see off him)
    e. .1sg focus on the subject of a transitive verb
       ex.: met ai.
          I   shot-φ-infl(V)1sg
         (it is I who shot)

  From these data alone, we can find the following interesting features in this language:

(i) Comparing the sentences (19a) and (19b), we find that the focus marking on the intransitive subject and that on the transitive object are identical. This implies that Yukagir has an ergative structure at least in focus marking6).
(ii) The sentences (19c) and (19d) shows that the focus marker on the predicate is identical both on intransitve and on transitive predicate.
(iii) The sentence (19e) has no morphological focus marker. Nagasaki (1995) explains that the transitve subject gets no focus marker even when it is focussed and the focussing is indicated byφ-ending of infl(V). However, if the obligatory coocurrence of the focus maker and the corresponding inflection paradigm is regarded as a structural principle, it has to be explained why it is not applied to this case. I doubt whether the absence of an inflectional paradigm can actually compensate the absense of a focus marker.
(iv) The problem (iii) is related to the question whether this language has sentences without focus. The answer is, of course, yes. In the text data, we find many sentences containing constituents without any focus marker. In such cases, many of them are specialized by an aspectual feature, e.g. with a durative suffix -nun-, the temporal specification for future -te-, or some other modalities. But it is not yet obvious how the focussing is related to these grammatical specifications. Nagasaki (1995) shows an interesting example of a focussed adverb without marking.

(20) question:mo-ni     “quanin    kewej-nu-k ?”
          ((he) said  when do you depart ?)
    answer:mo-d’eN   “eguojie    kewej-te-jeN”
          said-infl-I “tomorrow  depart-future-infl(I)1sg”
          (I said “I depart tomorrow”)

In this sentence, the adverb “tomorrow” is regarded the focus owing to the contextual information that it is the term questioned in the previous sentence. But as no focus marker like -leN, -ek can be attached to an adverb, it remains as a bare form. And the verb, which should have the prardigm infl(I), is used here without the focus marker me(r)-. She explains that, as the focus does not fall on the predicate, it has to lie necessarily on another terms than the predicate, in this case, on the adverb. But is it not simpler to explain these cases as focus-less sentences of this language, relying more on grammatical force of word order and the contextual informations of dialogues ?
(v) In connection with the problem (iv) above, it comes into question whether the inflection paradigms (I) - (V) have a grammatical force for indicating the focus. Have they powerful enough to produce a focus out of constituents without explicite focus marking? We saw above that the focussed subject/object coocurs with a corresponding paradigm and that, on the contrary, the latter can appear without any focussed constituent. The paradigms indicate, of course, the variety of persons, the transitive/intransitive opposition and the relative weight of itself among the constituents. The inflection types infl(I), infl(III), infl(IV) and infl(NEG) are complex enough to show oppositions among paradigmatic items, and on the contrary, the types infl(II) and infl(V) for the subject focus differentiate the oppositions only minimally. It seems reasonable, therefore, to assume that they does not function to indicate focus positions.
 The standard negation of Yukagir is indicated with a prefix el(e)-. When the predicate is focussed, it appears an alternative opposition: me(r)- markes the affirmative versus el(e)- the negative predicate focus, e.g. (Nagasaki (1995)),

(21) a. tudel mer-u:j.
      he  focus-go-infl(NEG)3sg.
      (went he did)
    b. tudel el-u:-φ.
　　　he   NEG-go-infl(NEG)3sg.
     (went he did not)

 In (21b) above, the negation involves the predicate, so that the predication itself is negated. If the scope of negation falls on another constituent, el- is prefixed to it, where the predicate gets the paradigm infl(NEG). Nagasaki (1995) shows the following examples:

(22) a. the subject is negated:.
      el-tudejlek      pun-nun-φ,           me-t’ale-n’e-j.
      NEG-he himself  kill-durative-infl(NEG)3sg focus-add-gerund-infl(I)3sg
     (it is not he who is killing, he has someone who helps him)
　    b. an oblique constituente is negated:
      el-met-in’      kelu-φ.
      NEG-I-directive  come-infl(NEG)3sg
     (it is not to me that (he) came)
    c. an adnominal verb is negated:
      met t’unNre-qane met ama: aruu [el-amutneN]  eju-φ.
      I   mind-(obj)  I   father word NEG-well   hit-infl(NEG)3sg
     (literally: my father’s words did not hit my mind well)

 As far as the examples above are concerned, the prefix el- has not only a predicative force of negation, but also it marks the scope of negation. It “quantifies” and marks a constituent negatively. But this double function of el- appears separately when the negative focus falls on a transitive object. Nagasaki (1995) writes:

(23) tudel ama-  t’e  -d -ile  -k      el’      -bun’-mele.
    he   good-gerund-d-deer-focus   NEG-focus-kill-infl(IV)
   where -d- is a connecting connsonant between vowels.
   (the good deer, he did not kill)

 In this sentence, the transitive object is marked with the affirmative focus marker -k, and the predicate is negated by el-, but the praradigm is the affirmative infl(IV). The focus marking, negation and paradigm type are used separatedly. If instead of infl(IV) -mele the negative ending for 3sg -φ appears like the following ungrammatical sentence, the correspondance between the negative focus and the predication paradigm would be preserved:

(23’) ?tudel ama-t’e-d-ile-k            el’-bun’-φ.
     he   good-gerund-d-deer-focus   NEG-kill-infl(neg)3sg.

The structure of the sentence (23) seems to be related somehow to the ergative construction of this language. But the ergativity does not appear in full, as Krejnovich (1982) notes, Yukagir is only quasi-ergative. As far as focus marking is concerned, it is found only in an affirmative structure. Here remain more problems for further consideration.
 In a great number of language data of Krejnovich, Kurilov, Endo and Nagasaki, there seems to exist no particular morphematic marker for the theme or the topic of a sentence. Of course, there are many sentences which begin with personal pronouns, temporal or local words in the leftmost position. If they have no focus marker attached, they can be interpreted as thematic sentences especially when the terms concerned are prementioned beforehand in the pre-context or in the assumption of the speech partners. We find scaresely mentions about a prosodic feature marking the theme/topic-marking in the literature. But it seems to be reasonable to assume that the leftmost term is somehow related to the spoken pre-context, so that it works just as theme. Let us observe some typical examples from Nagasaki (1997):

(24) a. mit niNie     juoka:-de-qa  aqal’we-jili.
      we stomach ache-locative  laugh-infl(I)1pl
      (we laughed till our side ache)
    b. taN  t'upt'e        wa:j qal:-t’e k0de   Nol-le'-n'.          tude dite ban-l'e-n'.  
       that chukchee too  terrible-adj  human be-like-infl(I)3sg  he  like be-infl(I)3sg
      (that Chukchee looks like a terrible man. he looks like that man)
    c. aduNn’e kin-ek     lajse-ri-mele?
      that   who-focus  bark-supplement-obj-infl(III).
      (to whom that barks?)

The loftmost NPs in three sentences appear to be the theme though they have no morphematic marker. Typical is the sentence (24c), which has focus on the transitive object. The lesftmost that seems to satisfy the contextual condition for the theme, but without any particular morphematic markers.
 We may conclude, therefore, that this language has no special morphologic means for a theme marking, perhaps, nor prosodically. To compensate this lack, it has a rich means for marking a focus as we see above.
 As generally assumed, the focus falls in the scope of negation. What is remarkable in this lanugage is that the marking of the scope of negation works parallel to the affirmative focussing, except for the transitive object like (23). Both markings are so dominant that it has a deep root in its structure of this lanugage.

 Summing up the focus marking of Yukagir we get the following map for the function and distribution:
(25)

verb class

subject

object

oblique

adverb

predicate

inflection type

  1

 



intr.

　○

　   I

  2

　×

　  NEG

  3

　○

　   II

  4

  ×

      NEG

  5

　φ

 　φ

　　 I

  6*

  ×

　　NEG

  7

  φ

　φ

　　 I

  8*

  ×

    NEG

  9

 



intr.

   tr.

　○

    III

 10

  ×

　　NEG

 11

　○

　　V

 12

  ×

    NEG

 13

　○

    IV

 14

　○

　×

    IV

 15

  φ

　φ

    III  

 16

　○

    NEG

 17?

 φ

　φ

    III

 18?

 φ

    NEG

Explanation of symbols in the table:
○: affirmative focus: -le(N), -(e)k, me(r)-
×: negative focus: le-
φ: zero-marking (only relevant positions are marked)
intr: intransitive predicate
tr: transitive predicate
*: data from Nagasaki (1995), which were not mentioned above
?: supposed because of no data

4. Speculations on Typology of Topic and Focus
 From the linguistic data and their analysis above, we will venture to speculate about some typological features of the topic and focus markings of Japanese and Yukagir in hope of finding a clue to know how this type of information processing is distributed in human languages.

4.1. Theme Language versus Focus Language
 We can distinguish two types of languages: A type has a morphematic theme marker as Japanese and another has focus markers as in Yukagir. In this theme type language exists no special morphematic elemtent indicating a focus. The contrary is also the case in a focus language, which has no special morphematic means to indicate a theme. There may exist languages which have distinctive morphematic markers for theme and for focus above. We have to search them. And languages without any morphematic marker for theme nor focus can be found easily.
Summerizing this rough typological distribution as follows:

(26) distribution of theme and focus languages

type

 theme marker

　focus marker

 example

 1

　　 ○

　　　×

 Japanese

 2

     ×

    　  ○

 Yukagir

 3

     ○

　　　○

   ？

 4

     ×

      　×　

  German

   where ○: exists, ×: lacks

4.2 Multifunction of Theme and Focus markers
 NP-wa in Japanese marks the theme of a sentence on the leftmost position. But it can be  scrambled from the position toward right. Difficulties arise when many NP-wa’s appear in a sentence. Among them, only the first one is regarded as the theme, and others are interpreted as contrastive. But a contrastive NP-wa can stand in the leftmost position, if the sentence contains only one NP-wa or if it moves to the position by scrambling. In this case the ambiguity thematic/contrastive appears.
 In negative sentences, contrastive NP-wa’s are involved in the scope of negation. They function as focus in this context. But in affirmative sentences they can be interpreted as focus, too, if they are specified with [+dp...]. In such a case a negative element can be found somewhere in the context.
 In sum, NP-wa is thematic in the leftmost position in the underlying structure before scrambling. The second and the following NP-wa are contrastive and function as focus when the contextual information [+dp...] is relevant. That a NP-wa plays a role of focus marker compensates the lack of a particular marker for focus in this language. It has double functions as theme or focus under different syntactic and contextual conditions.
 Yukagir has, on the contrary, no theme marker, but two types of fucus markers: an affirmative and a negative one. They can be attached to main sentence constituents: the subject, the object and the predicate among others; the negative focus marker can be attached an obligue NP, too, but they cannot coocur with adverbs. The focus positions correspond to the types of inflection paradigms. Namely, what type of inflection paradimgs occurs is determined according to what is focussed. But the correspondance are not perfectly obligatory because, of course, there are sentences without forcus. The inflection paradigms are differentiated primarily according to moods, i.e. for affirmative, negative, imperative, etc. Among them the affirmative mood alone is differentiated corresponding to focus positions. But it is not obvious whether inflection paradigms themselves have a grammatical force to indicate the focus position. If it were the case, an inflection paradigm alone could indicate what is focussed. Nagasaki (1995) seems to think about it positively.
 This language seems to have scrambling, too. An oblique constituent without focus and an adverb is moved often to the leftmost position. They can be regarded as the theme of the sentence even when another focussed constituent coourrs. Ellipsis also appears; especially when the focussed object moves to the leftmost position, the non-focussed subject is omitted.
 We can summarize the main morphosyntactic restrictions on NP-wa in Japanese and the focus marking of Yukagir as follows:

(27

 

   restrictions

 X-wa of Japanese

   focus markers in Yukagir

 affirmative

   negative

1.1

 in leftmost position

   thematic

         focus

1.2

 in other position

 contrastive=focus

    focus

    focus

2

 in the scope of NEG

    focus

     △

     ○

3

constituents concerned

   NP,VP, adv

 sbj.obj.pred

sbj.obj.oblique-  

  NP, pred,adv

4

  scrambling

theme:○, others　◇

　 　○

     ○

5

  in context [+dp _ ]

     focus

     focus

     focus

　　where ◯: the morpheme concerned appears,i.e.OK, ◇: possible, △: a special configuration  appears

    Note that Japanese wa is not mono-functional. Apart from our traditional view, it indicates either theme or focus according to different syntactic, locutional and situational contexts. If we want a single label for these functions, we have to choose the “selective” as Rickmeyer proposes. These different functions come to light explicitely in collaboration of the lexical meaning, syntax and the presupposed contextual information. Our lexicon in brain is “economic” enough to register only one morpheme wa which is realized in two ways in actual contexts. On the contrary, Yukagir has no thematic marker, but it can represent the thematic meaning by means of the syntactic condition that the constituent stands in the leftmost position of a sentence. That the language has different types focus marking for affermative and negative sentences seems to be somehow superfluous. If they coocurr obligatorily with the corresponding inflection paradigms, this inflectional variety seems to be an excesss, too. But if we think that the inflection paradigm for affirmative predicate focus is basic and, therefore, the others have far less distinctive oppositions, it can be attributed to a historical verbosity of this language. Perhaps these have some raison d’etres, too, which we do not know.  

4.3. Quasi-ergative Construction
 It has been asserted that the socalled emotion verbals with a “transitive meaning” like -ta-i, -hosi-i, suki-da, etc. require a special syntactic construction which has the “hyper-subject” and the embedded verb phrase in which the object is marked with a subject marker ga. e.g.

(28) [[hyper-sbj boku wa][[VP kanojo ga] suki-da]]
    (I like her)

Sentences like this are often interpreted as an ergative fragment of this language because of the subject marker ga added to the transitive object. If we accept this interpretation, we find here that the ergative construction in Japanese concerns the thematic NP-wa in the leftmost position. But the construction are very restricted to a verbal class above mentioned.
 Krejnovich (1982) mentioned “quasi-ergative” construction in Yukagir with the examples like our sentences (19a) and (19b). The construction is, according to him, also very restricted; it concerns only the identical focus marking for affirmative subject and for transitive object. But note that the both fragmental ergative constructions are related to the morphematic marking of focus. He is surely right to labelling this construction as “quasi-”ergative. As far as we see, there seems to be no reason to appreciate these constructions as relevant grammertical component of the languages. But it is interesting to see that the fragmental ergative structure appears in theme language Japanese with a theme marker, but in focus language Yukagir by means of the focus marker. But note that the markers appear in the distinct crucial places:

(29) a. in Japanese  TP                 b. in Yukagir   TP
          　　　　   ┌─┴─┐                       ┌─┴─┐
           　　　　　theme    VP           　　　                    VP
　　　　　　　　┌─┴─┐                        ┌─┴─┐
                                              　　　　　　　　　   focus

4.4. Marker in Contexts
 A small morpheme has often many meanings. We can say that it is ambiguous. And it is often said, too, that the ambiguity can be dissoved, if its meaning is described precisely enough to condense them into an abstract formulation. Surely, a semantic description may not remain just as an exhibition of particular usages in various contexts. An abstract meaning is actualized distinctively according to different context. In our case. wa means only “selection X out of X*”, but it function as the theme to indicate the intersentencial positioning of the sentence it precedes, or as the focus to indicate the intrasentencial weighting of the term it marks. They have different syntactic condition: the theme is placed in the leftmost position, the contrastive in the other. The lexical and contextual condition for the theme is [+m...], but the focussed term has a more heavy contextual condition [+dp...].
 Yukagir is a typical focus language. The focus marker is attached to main constituents of a sentence systematically except for the negative transitive object. But if we see the “exception” closer, we find that the opposition is represented simply:

(30)　　　　　　   negative VP　　　　　　　　cf.   affirmative VP
　　      　　　　            ┌──┴──┐             　　　　　　  ┌──┴──
　　　　　　　　　NPobj　　　  VPtr                           NPobj　　　　 VPtr
    　　　　　┌─┴─┐　 ┌─┴─┐          ┌─┴─┐　 ┌─┴─┐
　　　　　　N      FOC   FOC    　 V           N      FOC   FOC     V
 　                   |     　 |　        |                　　　　   |　　     |　　       |
                   -leN/-ek  　　　 le-   infl(IV)           -leN/-ek  φ    infl(IV)

If we observe the distribution of focus markers in the light of the negative-affirmative opposition as (30) above, we find that the focus markers are distributed for each pair of inflection paradigm in a simple opposition: afirmative versus negative construction. The morphematic focus marking of this language works on this principle. This collaborates with other morphosyntactic devices like the repertoires of various inflection paradigms. We have also seen that, if a non-focussed constitutent it is in the leftmost position, it looks like a theme. That means that the lack of focus marking can make the thematic interpretation possible if it stands in the leftmost position.

 Summing up the speculations above, we can imagine the pictures as follows:

(31) a. For NP-wa in Japanese as a thematic language:
      NP-wa has the meaning: selection X out of X
      (i) in the leftmost position ⇒ theme,
        if X is under the condition [+m ...]      
      (ii) in the second and following positions ⇒ focus
         where X must have the feature [+dp ...]
　    b. For focus marking in Yukagir as a focus language:
      (i) Focus markings stand in affirmative versus negative opposition,
      (ii) The focussing corelates with inflection paradigms, which have no semantic force to
         indicate a focus.
      (iii) Constituents without focus marking in the leftmost position is possibly ⇒ theme.

 A theme carries an information to connect the sentence leftwards in order to establish its inter-sentencial constellation. For this purpose, its leftmost positioning is crucial. In Japanese its scrambling is possible due to the overt morphematic marking wa. If a language lacks such an overt term, the constituent in the leftmost position alone can permit a thematic interpretation. The focussing is, on the contrary, a gramatical procedure to place a special informational weight onto a constituent within a sentence. It works for an intra-sentencial information processing. Yukagir shows that there can be an opposition: affirmative versus negation focussing. Surely, many languages have lexical means for representing sentence-internal focus, which has not directly to do with negation. As generally assumed, focussing conerns negation which presupposes a contextual features [+m ...] & [+dp ...]. On the base of this contextual presupposition, a negative predication is made onto the focus, so that a negative focussing appears. The notions, theme, focus and negation, are in mutual relationship, but they appear overtly under typologically different grammatical conditions.

トップへ