Take the 2-minute tour ×
Linguistics Stack Exchange is a question and answer site for professional linguists and others with an interest in linguistic research and theory. It's 100% free, no registration required.

If a language permits only CV or CCV syllables, how does one tell where the phonological word boundaries are?

For example, suppose a language has a definite morpheme, /li/ that conveys what English "the" conveys. Suppose this language has another morpheme, /stulu/, meaning "dog."

If these two morphemes occur in the order just mentioned, how can you tell whether we have a particle + noun (li stulu) vs. a single word with a definite prefix (listulu)?

I suspect that stress could be one way to differentiate these two, but are there other ways?

share|improve this question
4  
just curious, can you explain why it has something to do with the syllable pattern? –  Louis Rhys Mar 2 '12 at 18:02
1  
I thought that CVC, CCVC, CVCC, or CCVCC syllables would increase the chances of creating consonant sequences that violate phonotactic rules. In single compound words, such sequences could be eliminated through omission, substitution, or deletion. Such processes would not be needed if a word boundary broke up the forbidden sequence. So if Language X forbids “dkfp,” then we could have this form, which could be 2 words or 1: adkefpo. But “adk fpo” would definitely be separate words. –  James Grossmann Mar 3 '12 at 3:43
2  
Possibly, but not necessarily. There are languages in which any cluster that occurs between words can also occur word-internally. –  Dan Velleman Mar 3 '12 at 3:57
add comment

3 Answers

up vote 7 down vote accepted

Generally, you can't. Stress is a possible disambiguating strategy, but even in your hypothetical language, the two possibilities could have the same stress pattern. As with other forms of ambiguity, context is often sufficient to prefer one interpretation over another.

Syntactic constraints can play a large part: for instance, in your hypothetical language, li listulu could not possibly be analysed as *li li stulu, since, if li patterns with determiners cross-linguistically, the combination of two adjacent determiners would be blocked.

Interestingly, the English adder and apron were diachronically nadder and napron, the combination a napron having been re-analysed as an apron, showing the potential for fluidity in word boundaries through re-analysis.

share|improve this answer
    
li stulu and listulu can't have the same stress pattern. It's always one stress versus two. I don't think the part on li li stulu is clear. At least I didn't get it. –  kamil-s Mar 2 '12 at 7:45
    
@jogloran Tone can also disambiguate. –  Gaston Ümlaut Mar 2 '12 at 12:35
2  
@KamilS. I think that's incorrect. In English, and all other languages with determiners I can think of, determiners like an form a prosodic unit with their noun, and only receive stress in a marked reading. If li stulu is anything like an apple, then it would receive one stress only. –  jogloran Mar 2 '12 at 12:37
    
@jogloran If you're sure there's no difference in secondary stress then sorry, I'm taking it back. –  kamil-s Mar 2 '12 at 16:19
add comment

I don't quite see the link between your question about word boundaries and what syllables the language permits. Could you elaborate?

As for one versus two words, I don't think there is a universal method. It mostly depends on what the particular language does to its words and affixes, e.g.

  1. Stress is perhaps the most universal method, as almost all languages have it, and in your example it must be either one or two. Counterexamples: 1. one can always claim li is a separate word which might lose its accent in certain combinations; 2. stress solves nothing in cases such as Polish w 'in' or z 'from' which are only made up of a single consonant and so can't be stressed anyway (they count as separate words, in case you're interested).
  2. The possibility to insert a separate word in between can also help answer your question. How would you say 'the smart dog' in your language? Can 'smart' go between li and stulu? Counterexamples: German, Dutch, Hungarian &c. separable verbs are typically counted as one word, although they can be separated, their order swapped, and other words inserted in between.
  3. Vowel harmony is a good method in languages which have it. Counterexamples: Hungarian -kor temporalis does not harmonize with the stem, yet it is counted as a suffix rather than a separate word.
  4. Mutations which only happen on morpheme boundaries. Unfortunately, no ready example comes to my mind.
  5. Native speaker intuition would be a great method if it only weren't so shaky. It's a sort of chicken and egg problem with NS intuition and spelling, e.g. Russian v 'in' was first written together with the word, and later the spelling was changed to separate.
share|improve this answer
add comment

Good points have been made in other answers, but I feel that the notion of phonological word needs to be clarified. This term refers to a prosodic constituent, not a lexical one, so something that is a phonological word in one context might not be one in another. For example, the English word the can stand alone as a phonological word if it is spoken in isolation:

the

or if is under focus within a longer phrase:

Say "the" for me.

But in other contexts it might cliticize to a neighboring word, like a noun that follows it (as @jogloran mentions in a comment under the original question):

Eat the eggplant.

The fact that English orthography dictates that we put a space between the determiner the and the noun that follows it is just a convention that has no bearing on phonological word boundaries.

Since Japanese allows CV and a restricted set of CVC but not CCV syllables, it is not directly relevant to the OP's question (i.e. if one encounters a ...CCV... sequence, one can be confident that the first C is in the coda of the previous syllable). That being said, here's an attempt to address @Mechanical snail's question about Japanese:

The concept of the phonological word does not play a huge role in Japanese prosody. Pierrehumbert and Beckman (1988) note that there is some "segmental allophony" that can take place at the edges of the phonological word, but these phenomena are dependent on the combination of phonemes at those edges and may often not come into play. The accentual phrase is a much more crucial constituent when it comes to the domain of the lexical pitch accent and accentual phrase tones, so tonal information would help a Japanese speaker differentiate between syllables grouped into a single accentual phrase and those falling into two adjacent accentual phrases. Often an accentual phrase can consist of a single phonological word, in which case the two domains demarcate the same lexical material. A noun plus a particle can form such a constituent:

nasu-o

eggplant-ACCUSATIVE

This grouping of a noun and its particle into a single prosodic constituent is analogous to the determiner-noun grouping in English.

To bring everything back around to the OP's hypothetical language, if li is a definite article, and it behaves like the determiner the in English or the particle -o in Japanese, it will form a single prosodic constituent together with stulu whether it is a prefix or a standalone morpheme.

share|improve this answer
add comment

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.