Scale of homophony

[orros]

Dear colleagues,

I am looking for an article or website displaying information about research on homophones and homophony in different languages.

In particular, I am interested in finding a sort of scale of homophony meaning with that a graphic or numeric representation of the continuum going from languages with a few examples of homophones, to languages with many homophones such as mandarin Chinese.

I have searched a lot on the internet but I did not find anything of academic interest.

Thanks for you kind help,

Orros

 

[dihydrogen monoxide]

I would say that languages where phonetic puns are possible would have large homophony scale, like English and French.

 

[Ben Jamin]

Homophones are most common in languages in which the words are eroded in a high degree, for example in French.
Erosion is a process of words loosing their phonemes, both consonants and vowels, and becoming shorter, until almost nothing is left, as in French eau and août where only one vowel is left. Slavic languages have few homphones, and I will expect a very low number of homophones in Finnish and Hungarian.

 

[orros]

Thanks for you kind replies.
I am just a little bit surprised by the shortness of information about this topic on the internet. One may find a lot of research about homophony within a specific language but almost nothing about contrastive studies among languages. In comparison, when it comes to morphological topology, the abundance of information is impressive.
I think cross-linguistic homophony is quite an interesting topic, especially when trying to identify objective criteria aiming to characterize the difficulty of a language both for native speakers and foreign learners.
Regards,
Orros

 

[Penyafort]

I guess that the number of homophones will be proportional to the combination of two facts: a high number of monosyllabic lexical units plus a relatively short number of possible phoneme combinations. That is why Chinese is an extreme example and that's why "words" in modern Chinese are rather formed by two units instead of one.

Regarding the main Romance languages, I'd dare say, from more to less:

1 French​

2 Catalan​

3 Romanian​

4 Portuguese​

5 Spanish / Italian (I'm undecided here about which one would rank higher)​

 

[bearded]

and that's why "words" in modern Chinese are rather formed by two units instead of one.

Oh I thought that tones were sufficient to distinguish possible homophones in Chinese..

 

[AndrasBP]

Slavic languages have few homphones, and I will expect a very low number of homophones in Finnish and Hungarian.

Hungarian does have some homophones but of course much fewer than English or French.
I think Russian and Lithuanian have even fewer than Hungarian.

 

[Awwal12]

I think Russian and Lithuanian have even fewer than Hungarian.

It's hard to tell (homophones in Russian are certainly more widespread than full homonyms, due to the massive vowel reduction and some orthogrpagic peculiarities), but complex morphology (and the greater average length of words which comes with it) plus comparatively free phonotactics don't really favor the appearance of homonyms.

 

[Hulalessar]

I guess that the number of homophones will be proportional to the combination of two facts: a high number of monosyllabic lexical units plus a relatively short number of possible phoneme combinations.

Not necessarily. A language with a high number of possible unique syllables and/or high number of monosyllabic lexical units may still have a large number of homophones. Just because because a language has a high number of possible unique syllables does not mean it does not use some syllables much more than others and homophones do not have to be monosyllabic.

If you want to compare one language with another a dictionary definition of "homophone" is not really sufficient as you run in the problem of what amounts to a (separate) word. Some languages get close to the case that you cannot distinguish between a word and a sentence so that, for example, a noun cannot exist unless you say something about it. It can though be imagined that ambiguities are possible in such a language. In tonal languages are tones to be counted or discounted? In languages where verbs conjugate are two identical forms with different meanings (for example Spanish andaba which can be either first or third person singular) to be counted as different words? Do the number of homophones equate to the number of entries in a dictionary? Some entries may have quite different meanings whilst others may be quite close. How are we to decide which to count?

 

[Awwal12]

In languages where verbs conjugate are two identical forms with different meanings (for example Spanish andaba which can be either first or third person singular) to be counted as different words?

These obviously aren't different words. A more interesting case, though, is when the basic (e.g. nominative singular) form of one word coincides with a secondary form of another, cf. for example Russian nom.sg. лук /luk/ "onion; bow" and gen.pl. лук /luk/ "of saddle pommels (or cantels)", from лука /luká/.

 

[Zec]

Syncretism is a kind of homphony, but is usually treated separately from it. The distinction is basically: syncretism = two inflected forms of the same word sound the same, homophony = two different words sound the same. So andaba '-1sg' and andaba '-3sg' indeed count as the same word in this approach.

 

[apmoy70]

Modern Greek has a certain degree of homophony as a result of iotacism (=the fronting of -η-/-υ- and merging of their pronunciation with -ι- (iota, hence iotacism) into /i / and the monophthongization of the ancient diphthongs -ει-/-οι-/-υι- and the subsequent merging of their pronunciation into a simple /i / a long process that was completed by the 10th c. CE). Examples:
Pron: [eˈk͡sar.ti.si] (fem.) = «εξάρτιση»rigging (of a ship), «εξάρτηση», dependence, «εξάρτυση», equipment, outfit.
Pron: [ˈli.ma] = «λήμμα» (neut.), word, canonical form of a term (linguistics), «λύμα» (neut.), drainage, «λίμα» (fem.), file, rasp.
Pron: [ˈme.li] = «μέλι» (neut.), honey, «μέλη» (neut. nom pl.), limbs, body of members, «μέλλει» (3rd p. sing. Present indic.), it's coming to, it's about to.
Pron: [ˈti.çi] = «τύχη» (fem.), luck, «τείχη» (neut. nom. pl.), fortification, walls enclosing a town or city, «τοίχοι» (masc. nom. pl.), walls, «τύχει» (3rd p. sing. Present indic.), it happens.
Pron: [fiˈtɔ] = «φυτό» (neut.), plant, «φοιτώ» (1st p. sing. Present indic.), to study, attend.

 

[Penyafort]

Oh I thought that tones were sufficient to distinguish possible homophones in Chinese..

It may usually be, but even so it is a rather low number of possible outcomes. Modern Standard Mandarin has a rather low number of tones in comparison to other tonal languages.

Not necessarily. A language with a high number of possible unique syllables and/or high number of monosyllabic lexical units may still have a large number of homophones. Just because because a language has a high number of possible unique syllables does not mean it does not use some syllables much more than others and homophones do not have to be monosyllabic.

But that's cheating a little because a high number of possible but unused syllables is like not really having them in the language. I mean, if the homophone of a word is a word that almost no speakers know, the risk of confusion is nonexistent.

In languages where verbs conjugate are two identical forms with different meanings (for example Spanish andaba which can be either first or third person singular) to be counted as different words?

I'd say most speakers wouldn't see them as homophones. The meaning isn't really different -we're just talking about a change in person- and the origin is the same one. Quite different from a verbal form and an unrelated noun, like traje or tuvo/tubo.

 

[Hulalessar]

But that's cheating a little because a high number of possible but unused syllables is like not really having them in the language. I mean, if the homophone of a word is a word that almost no speakers know, the risk of confusion is nonexistent.

But English still has far more possible syllables than Chinese even if you discount rare combinations. Whilst that may lead one to think that Chinese is likely to have more homophones it does not follow because Chinese is not actually monosyllabic. Even if it were, we can still conceive of a language which has a high number of possible unique syllables but chooses not to use most of them and has developed lots of homophones matching the number in Chinese.

At a guess, I would say that many languages have more homophones than their native speakers realise. I was surprised when I read that the word with most entries in the OED is "set" with 430. You may not regard every entry as a different word, but it does have a very wide range of meanings.

The runner up tp "set" are:

• Run - 396 (definitions)
• Go - 368
• Take - 343
• Stand - 334
• Get - 289
• Turn - 288
• Put - 268
• Fall - 264
• Strike - 250

I'd say most speakers wouldn't see them as homophones. The meaning isn't really different -we're just talking about a change in person- and the origin is the same one. Quite different from a verbal form and an unrelated noun, like traje or tuvo/tubo.

I am inclined to agree. However, what about English where the same word can be different parts of speech as in "I like to butter my bread" v "I like to put butter on my bread"?

 

[Terio]

ɛ̃ Another aspect of the matter is how many words are homonyms.

For example, in French, we have :

[vɛ̃] :
vin : wine
vingt : twenty
vain : vain
(il) vint : (he) came
(il) vainc : (he) wins

[fɛ̃] :
fin : end
fin: fine
faim : hunger
(il) feint : (he) pretends

[vɛʀ] :
ver : worm
vers : towards
vers : verse
vert : green
vair : vair (fur)
verre : glass

[ʃɛʀ] :

cher : expensive
chaire : pulpit
chair : flesh, meat

etc.