Is it really impossible for two different languages to have the exact same set of phonemes?
I read that somewhere and wanted to know if it was true.
asked Feb 21 '19 at 23:58
-
3It is possible to have the same set of phonemes but have them occur in very different distributions. In fact, it is almost impossible to imagine two languages having the same phoneme systems, unless they were the same language to start with, because naming the phoneme is only the start. Phonemes have sets of allomorphs, which occur in specific contexts, and deform into other sounds in contact, in patterns that differ from individual to individual, let alone from language to language. As with anything alive, variation in language is the rule, not the exception. – jlawler Feb 22 '19 at 1:20
-
If there were a pair of candidate languages that did have exactly the same phonemes, it might be possible to argue that although the two languages might have different names, they were actually dialects of the same language. – Greg Lee Feb 22 '19 at 9:46
-
Jlawler, what did you mean with "have phonemes occur in different distributions" and why that would allow for two languages to share the same phoneme? – Duarte Alfonso Martin Feb 22 '19 at 12:02
-
2Of course it is possible. It's just very unlikely. There are so many possible phonemes. A nearby language (or dialect (or accent)) is almost defined by having only a very small of phonetic differences. It's gotta be a special brand of serendipity (i.e. extremely low probability) for a distant language to have somehow found it's way back to the the original language's exact set. – Mitch Feb 23 '19 at 2:34
-
What your definition of phoneme would be, @Mitch? – Duarte Alfonso Martin Feb 23 '19 at 11:31
An example of two languages with exactly the same phonemes is Zezuru and Karanga, which are Shona languages. The same can be said about Skagit, Snohomish and Puyallup. These can be called "dialects", or separate languages, depending on your criteria: they pass the test of mutual intelligibility, but there are various non-inventory differences. It does depend on how you define "the same" in terms of a phonemic system. Usually people mean "same segments with the same phonetic values", which is what allows inventories to be compared across languages. In substance-free theories, features for classifying sounds (i.e. the phonemes) are not phonetically based, so the idea that two sounds could be "the same" versus "different" is meaningless. So it would also depend on what your definition of "phoneme" is, and what the basis for comparison across languages is.
North Pare and South Pare have the same phonemes, except that the voiceless lingual fricative in NP is pronounced [θ] and in SP is is pronounced [s]. In terms of abstract analysis, they are not different and there is no reason in either language to say "This phoneme is [+strident] (or [-strindent]) in this language". If you define phoneme properties in abstract relational and functional terms, they are the same; if you define them in terms of pronunciation, they are different.
-
Truly, the case of whether two phonemes are different or the same comes down to the definition we give to this unit. With that in mind, what are the two definitions that you give in the end of your answer? – Duarte Alfonso Martin Feb 22 '19 at 11:59
-
The former refers to substance-free theories of phonemes and the latter refers to e.g. the SPE theory of features. – user6726 Feb 22 '19 at 15:46
-
Sorry, idk them, could you specify each definition? From now I say that I find a little strange the usage of phonetic descriptions for phonemes, since they're abstract. – Duarte Alfonso Martin Feb 23 '19 at 1:45
-
Abstract doesn't mean "lacking in phonetic properties", it is simply "draw away" from direct experience. Even phonetic properties are abstract. – user6726 Feb 23 '19 at 1:46
-
Can you tell why " If you define phoneme properties in abstract relational and functional terms, they are the same; if you define them in terms of pronunciation, they are different"? – Duarte Alfonso Martin Feb 23 '19 at 11:27
It depends what you mean by "the exact same".
Phonemes tend to be named with IPA characters, but they're really just algebraic symbols, representing mental objects that we can't poke and prod at directly. That's why you see symbols like /ḱ/ (PIE) and /ḥ/ (Semitic) and /ƀ/ (Germanic) and /ъ/ (Slavic) that aren't anywhere on the IPA chart. Alice Kober, who laid the groundwork for deciphering Linear B, famously used numbers for all the phonemes (like /2/) so as not to bias anything.
The upshot is that, while languages may have phonemes with the same symbol, they may not be pronounced at all alike. PIE and its daughter Proto-Germanic both have a /d/, for example, but it's pretty clear that the two sounded nothing alike. Or for a living example, English /r/ and Brazilian Portuguese /r/ tend to sound completely different: in English it's an alveolar approximant, while in Brazilian Portuguese it's sometimes a glottal fricative!
If you want two languages that have to have the same phoneme names/symbols, well, those symbols are arbitrary. All you have to do is pick languages with the same number of phonemes and rename them to match. Nothing's stopping you from using /$/ and /&/ as phoneme names, and non-IPA symbols can be neutral about how exactly the phonemes were realized.
On the other hand, if you want two languages that have exactly the same allophones and realizations of every phoneme, that's going to be impossible. Take two speakers of English from different areas, and I guarantee their phonetic realizations will be different in some way.
So when linguists say "English and Japanese both have /t/", they're making a compromise, saying that English and Japanese both have phonemes that are "close enough" in their realization: English /t/ can't be [ts], and Japanese /t/ can't be [ʔ], but they're vaguely similar enough to be called "the same".
-
I think the matter has been a little bit more complicated now. I know that there are multiple reasons to choose a symbol over other to represent a phoneme, but I thought one of the reasons was always frequency of occurrence. The fact 《r》, a symbol representative of a trill, is used in both English and Portuguese, despite the fact in those languages this phoneme is almost never realized as a trill, seemingly turns my suposition wrong. I mention about phonemic representation bc it seems a key part of your explanation for why it is possible for languages to have the same phonemes. – Duarte Alfonso Martin Feb 22 '19 at 11:04
-
@DuarteAlfonsoMartin It's common practice to use the IPA symbol of the most common realization, but this isn't universal: for example, Proto-Indo-Europeanists use
/k/for a sound that was probably uvular, just because nobody can agree on how exactly it was pronounced. Similarly, in my dialect, English/p/is almost always aspirated (even after/s/), but it's still written as/p/without an aspiration mark. Basically, frequency of occurrence is a good guideline to follow, but nothing stops you from using/1/and/2/for unknown phonemes like Kober did. – Draconis Feb 22 '19 at 15:42 -
I see; so to you there no equal phonemes between languages, there are just phonemes that differ greatly in their realization and those that don't. Before I call it a day, can you tell me how exactly using alphabetic letters would get in the way of the translation process conducted by Kober? – Duarte Alfonso Martin Feb 23 '19 at 1:48
-
@DuarteAlfonsoMartin Mostly that the language was utterly unknown: all she could determine were which phonemes were the same and which were different, with no information on realization. A lot of her colleagues got very set on particular realizations and it hindered the process of actually determining anything. – Draconis Feb 23 '19 at 2:42
-
Okay, she didn't know the realizations, but I still can't see why theorizing over them with letters would hinder her work. – Duarte Alfonso Martin Feb 23 '19 at 11:32
