Suppose you were listening in on someone’s conversation and you heard this:
Q: When did you eat the tuna?
A: Yesterday, I ate the tuna.
Doesn’t something just seem off about the answer? It seems like it would sound better if he said, “I ate the tuna yesterday,” with the adverb at the end instead of the start. But it’s the same words either way, so why does it matter? The problem is that in the answer as it is, the topic and focus are switched.
The topic of a sentence is, put simply, what the sentence is about. But of course, it wouldn’t be much use to have a sentence about something that the listener doesn’t know anything about in the first place, because then the listener would have nothing to relate the sentence to and the new information would have no use for him. So, for a sentence to be useful, there’s also an additional requirement that the topic of a sentence has to be something that both the listener and the speaker are already aware of. The focus of a sentence is the new information provided by the sentence. Take the last sentence of the paragraph before this one, for example: “The problem is that in the answer as it is, the topic and focus are switched.” Here, the topic is “the problem”; the fact that there is a problem was already made clear earlier in the paragraph, so this sentence takes that known fact and expounds upon it. The focus is, “that the topic and focus are switched.” This takes the idea of the problem and provides new information about it, saying exactly what the problem is. Focuses could range from a single word to pretty much everything in the clause besides the topic. You might have been taught in school that a sentence consists of a subject and a predicate, where the subject is what the sentence is about and the predicate is everything else. But “subject” really refers to the syntactic construction that’s ordinarily used for indicating the doer of an action; the topic is what the sentence is about. In fact, I think the reason people think in terms of subject and predicate is because English is a subject-prominent language, meaning that the subject is ordinarily assumed to be the topic by default. Although even in English there can be sentences with topics other than the subject. The alternative to subject-prominence is topic-prominence. Topic-prominent languages present a sentence in terms of topic-comment rather than subject-predicate. Japanese, for example, even has a separate word for indicating the topic of a sentence. One archetypal example of a Japanese sentence where the subject isn’t the topic is, “象は鼻が長い” zō-wa hana-ga nagai. The most idiomatic English translation would be, “Elephants have long noses,” but what’s really going on in the original sentence is more like, “Elephant [topic-marker], nose is long”; when people try to teach Japanese to English-speakers, they ordinarily try to emphasize that structure by translating it as, “As for elephants, their noses are long,” or “Speaking of elephants, their noses are long.” The problem with that approach is that the topic marker is used more often in Japanese than phrases like “as for…” or “speaking of…” in English. Really, there’s no perfect way to render a sentence like that idiomatically in English while still maintaining the original Japanese sentence structure; the languages are just too different. But that’s the sort of thing that makes languages interesting, I think.
Anyways, topics generally like to come at the beginning of a sentence. Focuses come later and usually get the strongest accent in the sentence. (No, I don’t mean that if someone is talking with a Southern accent, he’ll say the focus with his r’s curled back even farther than usual. I mean accent as in emphasis.) I’m pretty sure these trends are common across many if not all languages, not just English. I mentioned that English has ways of making something other than the subject the topic; generally, you do that by putting it at the beginning of the sentence. This is why the question and answer at the start of this post sounded wrong. The person who’s asking the question asks when the other person ate the tuna; he already knows that the other guy ate the tuna, he just wants to know when. But the answer, “Yesterday, I ate the tuna,” puts “yesterday” at the beginning, in topic position, and “I ate the tuna” later, in focus position. So he’s talking about yesterday, which is confusing enough because the questioner doesn’t know yet how yesterday is relevant to the question at hand, and then presenting as new information what the questioner already knew. Instead, he should put “yesterday” at the end of the sentence and stress it to make it the focus.
Questions are a special case. In a question, the focus, the new information, of the sentence is unknown; that’s the whole point of the question. So, in questions, the focus is replaced by a placeholder word (“who,” “what,” “where”), and the rest of the sentence is topical information. A good way to figure out what the main focus of a sentence is is to think, “If I were to say this in response to a question, what would the question have to be?” Of course, there could be multiple answers depending on where the stress is. For example, “I ate the tuna yesterday” could be said in response to either “What did you eat yesterday?” (“I ate the tuna yesterday”) or “When did you eat the tuna?” (“I ate the tuna yesterday“), because both “tuna” and “yesterday” come at the end of the sentence, where they can be construed as focuses. It’s the stress that indicates once and for all which is the main focus. One funny thing about English―well, I don’t know, maybe other languages do this too. But one thing that’s funny at least about English is that in questions, the topic and focus seem to completely switch places―so all of a sudden the focus comes at the beginning of the sentence, and the topic gets the strongest stress. I find, for example, that when I ask a question like “What did you eat yesterday?” the stress naturally seems to fall on “eat,” even though technically the focus is “what,” and “did you eat yesterday” (i.e. “you ate yesterday” in question mode) is topical information. Or, if I’m asking specifically about yesterday (notice that word “about”; in other words, yesterday=topic) as opposed to any other time, the stress naturally falls on “yesterday.” I’m pretty sure Japanese doesn’t do this; from what I’ve heard, in Japanese, the question-word seems to get the strongest accent. Although I’m not exactly fluent, so don’t take that as an eternal truth. But it makes sense because the question-word represents the focus. Also, since the topic indicates what specifically someone is asking about, which is kind of important for making sense of the question, English needs to have some way of indicating the topic in questions; so it commandeers the stress from the focus. But with Japanese, since there’s a built-in method of indicating topics, there’s no need to invert the focus-accenting rule. One funny thing about Japanese (funny from an English perspective, anyway) is that syntactically, questions are really treated exactly like declarative sentences, with no special word order or anything, so that a question like “どうすればいいでしょ？” (“If I do what it will probably be good?”) is perfectly normal.
My Personal Theory on Topic-Prominence
I was walking around between classes a few weeks ago, when I noticed something that struck me as pretty interesting…
The topic provides the “main” idea of the sentence, while the focus expounds on it. Similarly, the head in a dependency provides the “main” idea, while the dependent clarifies and modifies it. The dependent is to the head as the focus is to the topic.
So, that got me thinking. Ordinarily, dependency grammar says that the verb is the highest word in the dependency hierarchy. But maybe that only applies to subject-prominent languages. Maybe in topic-prominent languages, the highest word in the dependency hierarchy is actually the topic, and the main verb depends on it. Maybe “topic-prominence” is actually something more like “object-orientation”; in other words, the speakers of topic-prominent languages actually think more in terms of things than in terms of predicates.
But that’s just the personal little theory of an upstart college freshman who hasn’t even taken any classes in linguistics. And I don’t know how you’d test a hypothesis like that anyway.