Mermaid Construction in Ainu

There has been a need for me to study Ainu for a project that I am involved, and it has been such a joy to study the language. One grammatical fact that caught my eye in particular was evidential (?) constructions with ruwe ‘fact’, hawe ‘voice’, siri ‘land, appearance’, and humi ‘sound’ (there might be more). While the original semantics of these words are presumably nouns, they have undergone a semantic change into a more grammaticalized one, particularly one that has to do with evidentiality.

Examples

First, see the following examples (from 『中級アイヌ語 幌別』 published by the Foundation for Ainu Culture in 2011):

1. uhuynupuri anak pirka nupuri ne ruwe ne. / volcano TOP good mountain COP fact COP ‘The volcano (Mount Tarumae) is a pretty mountain.’
2. e=omkekar hawe ne yakun, to epitta e=sini yakun pirka. / 2SG.SUBJ=catch.a.cold voice COP COND day whole 2SG.SUBJ=rest COND good ‘If you have caught a cold, it is good for you to rest all day.’
3. tanto rera yupke siri ne kusu, hokanasmip kor wa oman. / today wind strong appearance COP CAUSAL coat have.IMP CVB go.IMP ‘Since it seems the wind is strong today, take a coat.’
4. nen ka ek humi ne na, soyne wa inkar. / who EMP come.3 sound COP PART go.out.IMP CVB see.IMP ‘It seems somebody has come, go out and see.’

Functions

What functions do these four nouns have? Apparently, according to the textbook, they express the source of the information and the belief of the speaker regarding the statement.

• ruwe ‘fact’: the speaker is certain about the statement. tan pet rehe mukapet ne ruwe ne. / this river name mukapet COP fact COP ‘This river’s name is mukapet (for sure).’
• hawe ‘voice’: the speaker’s statement is based on auditory information (hearsay). tan pet rehe mukapet ne hawe ne. / this river name mukapet COP voice COP ‘This river’s name is mukapet (they said).’
• siri ‘appearance’: the speaker’s statement is based on visual information. tan pet wakkaporo siri ne. / this river overflow appearance COP ‘This river is overflowing (I’m seeing it).’
• humi ‘sound’: the speaker’s statemnet is based on information through other bodily sensory systems. nen ka ek humi ne. / who EMP come.3 sound COP ‘Somebody came (I sensed it).’

Mermaid construction

Is this a noun, or is this some more highly grammaticalized morpheme? At a first glance, it looks like their morphosyntactic occurrence aligns with that of nouns, since it is typically followed by ne (copula). In addition, the modification by a preceding clause with no relativizer (so-called “gapless relativization”) is also a trait of nouns. If we treat it as a noun and translate the first sentence with the literal nominal meaning, it would be something like ‘(It is) a fact of the volcano being a pretty mountain.’ (which, of course, does not really make sense.) So, it definitely feels like it has undergone some semantic bleaching, or some grammaticalization.

In fact, this sort of grammaticalization, where it is syntactically behaving like a noun but semantically more functional (or abstract, or grammatical, or auxiliary-like), has been reported in a number of languages. This type of construction is dubbed as “Mermaid Construction” by Tasaku Tsunoda, but I prefer to call it “Nominal Raising Construction” or “Nominal Control Construction”, depending on the underlying syntactic structure.

Proposal

I argue that the constructions with the four Ainu nouns mentioned above (ruwe, hawe, siri, and humi) are instances of Nominal Raising Constructions, and they mainly contribute to evidential and mirative (“surprise”) semantics of a proposition.

Mirativity

ruwe, hawe, siri, and humi all have a mirative function.

• tan sipuskep ipeono ruwe! / this millet yield.much fact ‘This millet yields so much! (based on the fact)’
• pirka oruspe ne hawe! / good story COP voice ‘It is such a good story! (after listening to it)’
• pirkano a=nuye siri! / well IMPERS=draw appearance ‘It is drawn so well! (after looking at it)’
• tan sarampe sonno riten humi! / this cloth really soft sound ‘This cloth is so soft! (after touching it)’

Evidentiality

The evidential functions were already described above.

Focus in Wh-question

This function is somewhat similar to Japanese の (no). In Japanese, when asking a Wh-question, it is more natural to insert a nominalizer enclitic の (no) after the verb.

• naze iku=no=desu=ka? / why go=NMLZ=COP.POLITE=Q ‘Why do you go?’
• ?naze iku / why go

Similarly, Ainu uses the nouns described above in Wh-questions.

• toanpe kamuy ne ruwe? / that kamuy COP fact ‘Is that a kamuy?’ (or in Japanese: are=wa kamui=na=no=desu=ka?)
• nekon e=iki ruwe an? / how 2SG.SUBJ=do fact Q ‘How are you doing it?’ (or in Japanese: dou suru=no=desu=ka?)
• nen kar pe ne ruwe an? / who make.3 NMLZ COP fact Q ‘Who made it?’ (or in Japanese: dare=ga tsukut-ta mono=na=no=desu=ka?)
• hempak pe an hawe an? / how.many NMLZ exist voice Q ‘How many are there?’ (or in Japanese: ikutsu aru=no=desu=ka?)
• e=kor aca ney ta horipi eraman hawe an? / 2SG.SUBJ=have a=uncle where at dance learn voice Q ‘Where did your uncle learn to dance?’ (or in Japanese: anata=no ojisan=wa doko=de odori=wo oboe-ta=no=desu=ka?)

The last two examples do not seem to have anything to do with auditory information source; I do not know yet what are the semantic differences between these usages.