Tell me about Topic Prominence

[p-glyphs]

Hey all. Anyone here an expert in topic-prominent languages?

Is there any interesting link between phrase heads and the topic of sentences? (perhaps how they behave... or how they're treated... or just any noteworthy similarity you've observed)

While playing around with the noun phrase, I was hit with how phrase head are kinda like the topic of the phrase... while topic is the topic of the whole phrase... or even the conversation.

I'm wondering cause Umu's head final and topic final. I've been marking phrase heads when they're modified... but I just played around with marking the topic in the same way... like this

ja heru
two TOP/house
'the houses are two (in number)'

ja heru he
two TOP/house TOP/man
'the man, the houses are two (in number)'
'as far as the man is concerned, the houses are two (in number)'
(comment: two)(topic: house)(topic: man)

but I've been using this:

jal.lora he
two.CLS.house TOP/man
'the man has two houses'
(comment: two houses)(topic:man)

I've been researching topic prominence among isolating languages (mandarin, SE Asia), but I think I've a very indo-european concept of topic.

Also... since this new idea of mine now means there'd be multiple topics. Are there natlangs that use multiple topics and, if so, how?

Cheers guys and gals,

Andrew

[roninbodhisattva]

I think there's an analysis of topics in which they're the heads of a functional projection (topic phrase) that takes a CP as a complement, but I'm actually not really sure. I'm sure there are people who think that's bullshit and those that think it's the most obvious solution in the world. Who knows.

Also, I don't really get what you mean about phrasal heads being like topics.

[p-glyphs]

I don't really get what you mean about phrasal heads being like topics.

Let's say in a noun phrase... the head is the topic of that phrase.

"three blind mice"... mice is the topic of the (noun) phrase.
"three blind mice bought seeing eye dogs"... three blind mice is the topic the entire phrase.

Dose CP stand for complement phrase? If so then maybe I'm onto something.

I ask cause I don't know. Is the analysis you mention contentious? How come?

[roninbodhisattva]

Dose CP stand for complement phrase? If so then maybe I'm onto something.

Complementizer Phrase.

I ask cause I don't know. Is the analysis you mention contentious? How come?

I really don't know anything about it, just that I've read that people say they're some kind of topic phrase that's some kind of projection (TopP or some sort). The exact details for each language and what people think about it are probably varied and contentious or topic (hah) for debate. Who knows.

Let's say in a noun phrase... the head is the topic of that phrase.

"three blind mice"... mice is the topic of the (noun) phrase.
"three blind mice bought seeing eye dogs"... three blind mice is the topic the entire phrase.

I suppose. It's...conceptually similar.

[chris_notts]

"three blind mice"... mice is the topic of the (noun) phrase.
"three blind mice bought seeing eye dogs"... three blind mice is the topic the entire phrase.

I suppose. It's...conceptually similar.

Except that in most analyses I've seen a clause need not have a topic. "Three blind mice bought seeing eye dogs" seems most naturally interpreted as lacking a topic, because both subject and object are indefinite and non-generic. In general, topics tend to be at least identifiable to the listener, because if the rest of the clause is "about" the topic it generally helps if the listener can identify it.

[Salmoneus]

Yeah - I suppose you could assume that the mice were the topic because they're the subject, but really there's no clear topic in that instance. [As English isn't topic-prominent, there needn't be]

I also don't see the point of saying that 'mice' is the 'topic' in 'three blind mice'. What makes it the topic? You may never have mentioned the mice before. "The three blind ones crossed the street. Three blind mice". - 'Mice' clearly isn't a topic there!


topic FINAL language? Aren't they very very rare? I don't know, I could be exagerating that.

[zompist]

I'm pretty sure you can have a double topic in French:

Jean, il en a deux, des maisons.

--this in the somewhat artificial context where we are talking both about Jean and about his houses.

However, in Mandarin, in a sentence like

Xiang bizi chang.
Elephants have long noses.

you shouldn't think of this as two topics, but as a topic (xiang 'elephants') then a subject (bizi 'nose').

[chris_notts]

topic FINAL language? Aren't they very very rare? I don't know, I could be exagerating that.

I think a lot of languages have a split. Somewhat unpredictable topics tend to come first (if there is more than one potentially topical referent available). But languages often also have "afterthought" topics, where a mostly predictable topic is mentioned last.

[roninbodhisattva]

]I suppose. It's...conceptually similar.

Except that in most analyses I've seen a clause need not have a topic.

This is exactly what I was looking for.

[Salmoneus]

topic FINAL language? Aren't they very very rare? I don't know, I could be exagerating that.

I think a lot of languages have a split. Somewhat unpredictable topics tend to come first (if there is more than one potentially topical referent available). But languages often also have "afterthought" topics, where a mostly predictable topic is mentioned last.

Oh, I know that - English is a great example, after all. But that's topic-final sentences, not topic-final languages. It's not the default word order.

[p-glyphs]

topic FINAL language? Aren't they very very rare? I don't know, I could be exagerating that.

I chose it to be rare. The languages I know use a lot of topic-comment constructions, which of course are always put the topic first. I swapped it around, recalling these very "afterthought" topics situations.

you shouldn't think of this as two topics, but as a topic (xiang 'elephants') then a subject (bizi 'nose').

I've read this in Li and Thomson's description of Mandarin. It's been really helpful in exploring this. The problem I find is that they mainly use examples of verbless sentences to describe topic... cause I guess when there's a verb needing an argument around... topic flies out the window. At least in my indo-european head.

The example from French is really interesting.

I also don't see the point of saying that 'mice' is the 'topic' in 'three blind mice'. What makes it the topic?

hmmm. Perhaps it would have been better to say that 'mice' is the head of the NP and 'three blind mice' is the "head" of the clause. Maybe I wasn't equating heads with topics but topics with heads... which is why I wanted to mark them similarly. Do you understand?

"Three blind mice, they bought seeing eye dogs."

[roninbodhisattva]

erhaps it would have been better to say that 'mice' is the head of the NP and 'three blind mice' is the "head" of the clause.

But "three blind mice" is not the head of the clause. Infl is the head of the clause.

[Yng]

I've read this in Li and Thomson's description of Mandarin. It's been really helpful in exploring this. The problem I find is that they mainly use examples of verbless sentences to describe topic... cause I guess when there's a verb needing an argument around... topic flies out the window. At least in my indo-european head.

No. For a start, Mandarin has no verb agreement.

[p-glyphs]

But "three blind mice" is not the head of the clause. Infl is the head of the clause.

Infl Can you tell me that this mean? Please

No. For a start, Mandarin has no verb agreement.

Uh huh.... but it does assign arguments. Otherwise it would be terribly *ambiguous*.

There are analyses that map a 'double subject' construction, and then topic-prominence rebuttals which say topic has more grammatical weight than subjects—like the topic-subject example zompist gave.

I want my language to have a 'topic-patient' feel

[Yng]

Uh huh.... but it does assign arguments. Otherwise it would be terribly *ambiguous*.

There are analyses that map a 'double subject' construction, and then topic-prominence rebuttals which say topic has more grammatical weight than subjects—like the topic-subject example zompist gave.

I want my language to have a 'topic-patient' feel

Except you said 'agreement'.

Also, in a hypothetical SOV language (Mandarin has some SOV traits but I can't remember if this is possible - it certainly is in some langs though), you could have a sentence 'dog cat eat' where 'dog' is the topic and the object, or where 'dog' is the topic and the subject. These things happen!

Incidentally, ambiguity is not as bad as people often think. Everyone goes through an 'oh no ambiguity' stage, I'm sure - but ambiguity runs rife in actual languages.

[roninbodhisattva]

Infl? Can you tell me what this means? Please

It's the head of the of a functional projection called the IP, the inflectional phrase, which is taken to be synonymous with the clause in the generative framework. It's where inflection like agreement, tense, aspect, etc. is located.

[chris_notts]

Infl? Can you tell me what this means? Please

It's the head of the of a functional projection called the IP, the inflectional phrase, which is taken to be synonymous with the clause in the generative framework. It's where inflection like agreement, tense, aspect, etc. is located.

As you say, this is theory specific. There are plenty of theories which would disagree about what the head of the clause is, and some of them even about whether clause heads (or heads in general) exist.

[p-glyphs]

Except you said 'agreement'.

Nope... it says 'argument'

Also, in a hypothetical SOV language (Mandarin has some SOV traits but I can't remember if this is possible - it certainly is in some langs though), you could have a sentence 'dog cat eat' where 'dog' is the topic and the object, or where 'dog' is the topic and the subject. These things happen!

You're right.

shu wo mai le
book I buy LE
'the book, I bought it' (topic/contrastive)

wo shu mai le
I buy book LE
'I bought the book' (contrastive

Incidentally, ambiguity is not as bad as people often think. Everyone goes through an 'oh no ambiguity' stage, I'm sure - but ambiguity runs rife in actual languages.

I agree. Those were my 'ironic' stars.

[p-glyphs]

Infl? Can you tell me what this means? Please

It's the head of the of a functional projection called the IP, the inflectional phrase, which is taken to be synonymous with the clause in the generative framework. It's where inflection like agreement, tense, aspect, etc. is located.

As you say, this is theory specific. There are plenty of theories which would disagree about what the head of the clause is, and some of them even about whether clause heads (or heads in general) exist.

Wow. Thank you for that link. I think this explanation is very amazing.

So the mechanism that changes buy to bought is actually the clause head.

Has this idea been tested on a language like Mandarin or Cambodian? (isolating, non-inflecting). Are then all verbs 'light verbs'/all VP's then vP's? What are the consequences of this analysis on topics?

[roninbodhisattva]

Infl? Can you tell me what this means? Please

It's the head of the of a functional projection called the IP, the inflectional phrase, which is taken to be synonymous with the clause in the generative framework. It's where inflection like agreement, tense, aspect, etc. is located.

As you say, this is theory specific. There are plenty of theories which would disagree about what the head of the clause is, and some of them even about whether clause heads (or heads in general) exist.

Yeah, that's the problem. I don't know of any theory that takes clause heads to exist while holding that the subject (or topic) is that head, do you? (If it there is such a one I would love to read about it, actually...)

Has this idea been tested on a language like Mandarin or Cambodian? (isolating, non-inflecting). Are then all verbs 'light verbs'/all VP's then vP's? What are the consequences of this analysis on topics?

Well, Mandarin and Cambodian would still be assumed to have IPs even though they don't have inflection, but I have no idea how those analyses work. But I think the main verbs in those languages would still be V, not vP.

And I have no idea about the topic prominence literature so I can't give you an answer on that. Look it up,

[Zhen Lin]

Question: Is there a topic-prominent language where verbs agree with the topic, rather than the subject? The only topic-prominent language I know with verbal inflection (Japanese) does not have concord...

[roninbodhisattva]

http://en.wikipedia.org/wiki/Topic-prominent_language

[p-glyphs]

Question: Is there a topic-prominent language where verbs agree with the topic, rather than the subject? The only topic-prominent language I know with verbal inflection (Japanese) does not have concord...

From my limited understanding, topics don't need to have any semantic relationship with the verb. The waters are muddied, however, because on many occasions they do. I think this what distinguishes topic-prominent languages from others languages.

Cool that Japanese 'proves' this with it's inflections, since so many topic-prominent languages (≈ topic-prominence descriptions) are isolating.

[merijn]

In GB (mainstream generative grammar from the 80's) the topic was usually a specifier of the CP. From the 90's onwards the topic is usually understood to be the specifier of a projection called TopicP (sometimes abbreviated to TopP).

++++

I think the idea that there are heads has been around since the 1930's, and I am sure that it predates Generative Grammar. However, what a head is differs from framework to framework, and is at the moment quite a hot topic in Mainstream Generative Grammar. It is also perfectly possible for a framework not to have the notion head, and still be generative in the sense that it generates grammatical sentences (Categorial Grammar is an example).
In the 90's there was the introduction of a great number of functional categories replacing traditional IP and CP. Some people have proposed up to 12 functional categories dominating VP. Anyway, a modest example of such an "exploded" IP and CP is the following structure: [ TopP [NegP [TP [AspP [vP [VP]]]]]]. In proposing such an elaborate structure for verbs (and similar elaborate structures for what used to be regarded DP's) some people have thought that there should be a special status for the lexical projection that is dominated by all those functional categories, in our case the verb, and call it the lexical head, and the layers of functional projections the extended projections of the lexical head. In some other frameworks that have the notion head, the head corresponds to the lexical head rather than the functional head of the structure, so in our example the verb would be the head. However, I have never come across a framework in which the topic would be the head.

+++

Quite a few Bantu languages have constructions where the verbs shows agreement with the topic rather than the subject. If I had any time I would look for a real example, but I think in Swahili you can say 7.book sa7.read 1.boy, where the numbers are the noun class markers and sa stands for subject agreement, where the sentence means "the boy reads the book" where the book is a topic.

[Salmoneus]

Infl? Can you tell me what this means? Please

It's the head of the of a functional projection called the IP, the inflectional phrase, which is taken to be synonymous with the clause in the generative framework. It's where inflection like agreement, tense, aspect, etc. is located.

I believe this location is found just to the left of the platonic form of the horse and just to the right of the quintessance of the colour red. You know, behind the Seventeenth Intellect of the Sphere of Jupiter, next to the world where triangles have four sides, right across the the sea of the meinongian potentia...