Australian Linguistics Society
Japanese indefinite nouns and gradation of non-specificity
Harumi Moore, Faculty of Asian Studies, The Australian National University, Canberra, Harumi.Moore@anu.edu.au
This paper argues that in spite of the term used to refer to them, indefinite nouns in Japanese express non-specificity and that their use is pragmatically motivated. The paper also shows that their different syntactic positions reflect different degrees of non-specificity.
English regularly marks the definite vs. indefinite distinction. By contrast, the Japanese noun phrase does not mark this distinction as regularly, although it can often be understood from real world knowledge, the immediate discourse or linguistic context, or from the use of demonstratives such as sono 'that' or words such as aru 'a certain'. However, there is a distinct grammatical class of words which are called "indefinites" or "indefinite nouns" in Japanese consisting of words such as nanika 'something' or dareka 'somebody'. These words are composed of an interrogative noun such as nani 'what' and dare 'who' and the suffix -ka. Backhouse (1993) states that they indicate a general category ('thing', 'place') but leave identity unspecified. I am going to examine their syntactic and semantic characteristics below.
2. Syntactic distribution of indefinite nouns in Japanese
Unlike English indefinite expressions such as some and any as articles and something and anything as pronouns, Japanese indefinite nouns can appear in three distinct syntactic positions. Firstly, they occur as the heads of noun phrases. (The linked examples below are taken from English novels and their published Japanese translations. Translations into English in single quotation marks are the author's)
(1a) Sutoobu-no naka-de maki-ga mishimishi to stove-GEN inside-LOC log-NOM crackling QUO oto-o tatete-ita. Tsui sakki made sound-ACC making-was Very a short while ago until dareka-ga ita no da. (SS) INDEF-NOM existed NMLZ COP 'Lit.: Wood was making a crackling sound in the stove. Someone was here until a short while ago.' (1b) Sophie could hear wood crackling in the old stove. Someone had been here very recently. (SW) (2a) Takkaa-wa nanika-o saguridasoo to shite-iru T.-TOP INDEF-ACC look for-VOL QUO doing-is yoo da. (SK) like COP 'Lit.: Tucker seems to be looking for something.' (2b) I could tell he was fishing for something. (FPF)
As the translations above show Japanese indefinites appear to roughly correspond to English compound pronouns someone and something in these cases.
As well as appearing as heads of noun phases, Japanese indefinite nouns also co-occur with a noun phrase:
(3a) Koko-ni nanika yakuhin-ga tsuite-iru na. (SK) here-LOC INDEF chemical-NOM attaching-is SP 'Lit.: There is some kind of chemical residue [on the corpse] here.' (3b) She's got some chemical here. (FPF) (4a) Dareka hoka-no hito-ga motte-iru kanoosei mo INDEF another-GEN person-NOM having-is possibility also aru wa. (FPF) exist SP 'Lit.: There is also a possibility that some other person has it.' (4b) Someone else could have it. (FPF)
In these examples Japanese indefinites seem to roughly correspond to English indefinite articles. Note, however, that although English indefinite articles appears pre-nominally, the Japanese indefinite can also occur after the NP as in (3a)' or further away from the NP with an adjunct between the indefinite noun and the NP as in (3a)":
(3a)' Koko ni yakuhin ga nanika tsuite-iru na. (3a)'' Nanika koko ni yakuhin ga tsuite-iru na.
Finally, indefinite nouns in Japanese can occur by themselves:
(5a) Moshi anta-no mi-ni nanika okot-tara ore if you-GEN body-LOC INDEF occur-if I donna ki-ga suru to omou. what kind of feeling-NOM do QUO think 'Lit.: If something happens to you, how do you think I will feel?' (5b) You know how I'll feel if something happens to you? (FPF) (6a) Kitto nijup-pun mo are-ba dareka kite-kureru probably twenty-mins about exist-if INDEF coming-give deshoo. (SI) probably 'Lit.: If there is twenty minutes or so, surely somebody might come here for us.' (6b) I can probably get somebody here in twenty minutes. (CD)
Another syntactic characteristic of Japanese indefnite nouns is that, unlike English compound indefinite pronouns which can occur as NP heads with post-modification, as in somebody I know and something to read, indefinite nouns in Japanese are rarely found in NP head position when they are restrictively modified. Instead, they occur outside the NP, and lexically general nouns such as mono '(concrete) thing' or koto '(abstract) thing', hito 'person' or tokoro 'place' form the head noun of the modified noun phrase:
(7a) Nanika yomu mono-o o-mochi deshita ne. INDEF read thing-ACC HON-have COP-PAST SP (SK) 'Lit.: He had some kind of reading material.' (7b) He always had something to read. (FPF) (8a) Kono atari-de kyuujitsu-ni mo renraku-ga this area-LOC holiday-LOC also contact-NOM toreru hito-o dareka shir-anai. (SI) take-POT person-ACC INDEF know-NEG 'Lit.: Do you know any person [at all] around here who can be contacted even on holidays?' (8b) Do you know anyone around here we might be able to reach on a holiday. (CD)
In (7a), the head noun mono 'thing' and its modifier yomu 'to read' constitute a noun phrase, while the indefinite noun nanika precedes the noun phrase. In (8a), hito 'person' is modified by kyuujitu ni mo renraku ga toreru 'who can be contacted even on holidays', while the indefinite noun dareka follows the noun phrase.
3. Japanese indefinite nouns and non-specificity
In this section I will point out that indefinite nouns co-occurring with noun phrases in Japanese express the non-specificity of the referent and not indefiniteness, and that their uses are pragmatically motivated.
I have shown above that indefinite nouns in Japanese occur outside the NP where the latter is restrictively modified. However, many examples were found in texts where English compound indefinite pronouns with modification are simply translated by a noun phrase with a head noun such as hito or mono but with no indefinite noun in Japanese:
(9a) Demo, kimi-ni mite-hoshii mono-ga aru n da. (SK) but you-DAT see-want thing-NOM exist NMLZ COP 'Lit.: There is a thing that I want you to look at.' (9b) But there's something you need to see. (FPF) (10a) Watashi-ni mise-tai mono-ga aru I-DAT show-DES thing-NOM exist n daroo. (SK) NMLZ probably 'Lit.: I think there is a thing you want to show me.' (10b) I take it you've got something for me to look at. (FPF)
Both examples mention something which one person wants the other person to see. Interestingly, an indefinite noun can only be added to (10a):
(9a)' Demo, kimi ni mite-hoshii mono ga *nanika aru n da. (10a)' Watashi ni mise-tai mono ga nanika aru n daroo.
Such a discrepancy occurs because Japanese indefinite nouns operate in a different semantic domain from English indefinite expressions. Givon (1993) states that among the indefinite expressions in English, any and no are non-referring (non-specific) but some can be referring (specific) or non-referring (non-specific). Lyons (1977: 455) also points out the specific and non-specific interpretations of some. The English compound indefinite pronoun something in the examples above expresses indefiniteness regardless of whether the referent is specific as in (9b), where the speaker himself knows what he is referring to, or non-specific as in (10b), where the speaker has yet to identify what he is referring to. In contrast, Japanese indefinite nouns can only be used with non-specific referents. This is why nanika can only occur in (10a)'.
Let us argue this point further. The following pair of English sentences show contrasting specific and non-specific interpretations of something:
(11) I had something nice for lunch today. (12) I want to eat something nice for lunch today.
The reading of something nice in (11) is typically specific as the speaker knows what he is referring to. However, the reading of something nice in (12) is typically non-specific as the speaker does not know what he is referring to. Fodor (1970) argues that specificity is not an inherent property of the noun phrase in English but is induced by linguistic contexts such as tense. The past tense is associated with actuality, verifiability and thus specificity of the entity referred to. Thus something nice in (11) tends to evoke a specific reading as the action of eating has already taken place. A non-specific reading of (11) is only possible when the speaker does not know what he was eating but knows it was delicious. On the other hand, something nice in (12) does not evoke any specific reference. The desiderative mood want to eat in (12) is non-factual and thus induces non-specificity. The speaker wants to eat any old thing as long as it is nice.
In Japanese an indefinite noun nanika can only appear in the desiderative sentence where oishii mono refers to a non-specific referent:
(13) Kyoo ohiru-ni nanika oishii mono-ga today lunch-LOC INDEF delicious thing-NOM tabe-tai. eat-DES 'Lit. I want to eat some kind of nice thing for lunch today.' 'I want to eat something nice for lunch today.'
Thus, Japanese indefinite nouns co-occur with non-specific noun phrases, although, as noted above, they are not obligatory.
If it is the case that indefinite nouns are not obligatory with non-specific noun phrases, it is important to determine when they do occur. Alfonso (1971: 798) states that "a sentence is sufficiently clear and meaningful without the interrogative word +KA" and that the addition of adverbial indefinites is "somewhat emphatic". In other words it is the speaker's motivation to emphasise non-specificity that decides the use of indefinite nouns. The following example demonstrates this point very clearly. The situation is that the speaker knows that he had a specific reason to climb the Tokyo Tower but he can no longer remember it, as he is now extremely frightened by the height. Therefore nanika is used to emphasise the non-specificity of the forgotten "reason":
(14) Are, ore-wa dooshite konna koto-o hajimeta Whoops, I-TOP why like this thing-ACC began n dak-ke. Nanika riyuu-ga atta NMLZ COP-I wonder INDEF reason-NOM existed hazu da. (I) expectation COP 'Whoops, why have I begun to do a thing like this? I know I had some reason for it.
A strong sense of non-specificity is always present when the indefinite noun is added to a noun phrase. This can be seen in the examples which we looked at earlier:
(3a) Koko-ni nanika yakuhin-ga tsuite-iru na. (SK) here-LOC INDEF chemical-NOM attaching-is SP 'There is some kind of chemical residue [on the corpse] here.' (4a) Dareka hoka-no hito-ga motte-iru kanoosei mo INDEF another-GEN person-NOM having-is possibility also aru wa. (FPF) exist SP 'There is also a possibility that some other person has it.' (7a) Nanika yomu mono-o o-mochi deshita ne. INDEF read thing-ACC HON-have COP-PAST SP (SK) 'He had some kind of reading material.' (8a) Kono atari-de kyuujitsu-ni mo renraku-ga this area-LOC holiday-LOC also contact-NOM toreru hito-o dareka shir-anai. (SI) take-POT person-ACC INDEF know-NEG 'Do you know any person [at all] around here who can be contacted even on holidays?'
The non-specificity of yakuhin 'chemical' in (3a) is expressed with nanika as if to say "some chemical was found but at the moment there is no clue as to what kind of chemical it is". The non-specificity of hoka no hito 'a person other than the person in question' is emphasised in (4a) with dareka as if to say "I think a person may have it, although I have no idea who it could be". Nanika yomu mono 'some kind of reading material' in (7a) seems to emphasise the non-specificity of the entity, possibly because the speaker, the restaurant owner, was not paying attention to what the customer had at that time. Dareka in (8a) emphasises that the speaker does not have any particular person in mind as long as he/she can be contacted on holidays.
The above examination has clarified that Japanese indefinite nouns express non-specificity, and that they can be viewed not as a grammatical element in the sense of being obligatory, but as elements used to achieve the particular pragmatic function of emphasising the non-specificity of the referent.
4. Degree of non-specificity and the syntactic positions of indefinite nouns
I will argue below that, as far as noun phrases in the nominative and accusative case are concerned, whether the indefinite noun appears as an NP head or whether it occurs by itself reflect different pragmatic motivations. We will connect this argument with Givon's (1993) observation that in English there is a gradation in the domain of indefiniteness, and that such a gradation is expressed by different linguistic devices.
Givon (1993: 224) argues that indefiniteness may not only be categorised in terms of referring (specific) and non-referring (non-specific) but that "there is a continuum of referential intent" (Givon: 224-5). To support his argument, he shows that English grammar codes this continuum by combining three devices: (a) the choice of indefinite article, (b) the presence of restrictive modification and (c) the specific nature of the noun. He also argues that the underlying factors reflected in the gradation of indefinite reference exist in the (a) psychological and (b) probabilistic dimensions. In other words, "How strongly does the speaker intend to suggest that they are referring to a particular individual?" and "What is the probability that the individual the speaker referred to is a specific individual?". In Japanese the degree of non-specificity seems to be reflected by whether the indefinite appears as a NP head with a case particle (lower in non-specificity) or whether it appears by itself (higher in non-specificity).
Let us examine the degree of non-specificity denoted by the indefinite nouns. First let us re-cite the examples in which indefinite nouns appear as NP heads:
(1a) Sutoobu-no naka-de maki-ga mishimishi to stove-GEN inside-LOC log-NOM crackling QUO oto-o tatete-ita. Tsui sakki made sound-ACC making-was Very a short while ago until dareka-ga ita no da. (SS) INDEF-NOM existed NMLZ COP 'Wood was making a crackling sound in the stove. Someone was here until a short while ago.' (2a) Takkaa-wa nanika-o saguridasoo to shite-iru T.-TOP INDEF-ACC look for-VOL QUO doing-is yoo da. (SK) like COP 'Tucker seems to be looking for something.'
(1a) refers to a past event and to someone that actually existed. Dareka followed by the nominative case particle clearly illustrates the conviction that, although the identity was unknown, there existed a human being who had been there until a short time before; thus non-specificity is low. In (2a) nanika followed by the accusative case particle seems to evoke the impression that, although the speaker does not know what the person is looking for, the speaker strongly suspects that the person does. Thus, again, the degree of non-specificity is low.
Now let us re-cite the examples in which indefinite nouns occur by themselves:
(5a) Moshi anta-no mi-ni nanika okot-tara ore if you-GEN body-LOC INDEF occur-if I donna ki-ga suru to omou. what kind of feeling-NOM do QUO think 'If something happens to you, how do you think I will feel?' (6a) Kitto nijup-pun mo are-ba dareka kite-kureru probably twenty-mins about exist-if INDEF coming-give deshoo. (SI) probably 'If there is twenty minutes or so, surely somebody might come here for us.'
The referents in the above examples are in the irrealis mood. In (5a) nanika appears in a hypothetical clause and refers to an event which may or may not occur. The speaker has no particular idea about what kind of event it might be. In (6a) dareka appears in a clause referring to a future event and refers to anyone who may be able to come to the scene. Both examples are therefore highly non-specific.
We can support this speculation - that indefinite nouns appearing as NP heads reflect a lower degree of non-specificity than those appearing by themselves - in a wider linguistic context. The first supporting evidence is the syntactic distribution of negative indefinite nouns. Such nouns only occur by themselves and they do not form NP heads:
(15) Daremo kon-akat-ta. INDEF come-NEG-PAST 'Nobody came.' (15)' *Daremo-ga kon-akat-ta. INDEF-NOM come-NEG-PAST
We can attribute this phenomenon to the higher degree of non-specificity of the referent, as referents in a negative sentence are non-factual, i.e. they do not exist.
The second piece of evidence concerns modification. As mentioned earlier, Japanese indefinite nouns are rarely restrictively modified. We can argue that this is because restrictive modification, as Givon (1993) says, increases the degree of specificity. However, when it occurs, such modification is only found with indefinite nouns in the NP head position:
(16a) Soreja, ERF-no dareka-ga saisho ni sono koto-o then, ERF-GEN INDEF-NOM first COP that thing-ACC kare-ni hanashita no ne. (SK) he-DAT told NMLZ SP 'Lit.: Then someone from ERF told him about that first.' (16b) Then, someone from ERF brought it to his attention first. (FPF)
We can attribute this phenomenon to the lower degree of non-specificity of the referent expressed by the indefnite noun in the NP head position.
The above examination supports our hypothesis that the degree of non-specificity can be reflected in different syntactic environments.
This paper has argued that Japanese indefinite nouns express non-specificity and that their co-occurrence with nouns is pragmatically motivated. They are not obligatory, but their presence emphasises the non-specificity of the referent. This paper has also shown that their different syntactic positions reflect different degrees of non-specificity. This phenomenon in Japanese is similar to Givon's (1993) observation that a gradation in the domain of indefiniteness is expressed by different linguistic devices in English.