Talk:Minimal pair

From Wikipedia, the free encyclopedia

Contents

[edit] Old talk

This article seems inconsistant about when to use [brackets] and when to use /slashes/. My psycholinguistics book says to use brackets for phones and and slashes for phonemes. Is there any reason not to follow this convention? --Ryguasu 22:09 Feb 28, 2003 (UTC)

English "let" + "lit" proves that phones /e/ and /i/ do in fact represent distinct phonemes [e] and [i].

I don't think these are the right symbols, at least if we're using IPA. The vowel in "let" should be /ε/, and that in "lit" should be /I/, no? Josh Cherry 00:55, 20 Oct 2003 (UTC)

That should be small capital I; otherwise you are correct. /e/ and /i/ distinguish "late" and "leet". -phma
Depends on what set of symbols you've chosen. Some sources for English choose "e", some choose "ɛ". Some choose "i" and "ɪ", some choose "iː" and "i", some choose "iː" and "ɪ". There's no universal agreement for how to use IPA for English. Hippietrail 07:13, 9 Mar 2004 (UTC)


but Wikipedia articles should be consistent with the International Phonetic Alphabet for English article (though that should mention such disagreement). Joestynes 12:20, 8 May 2005 (UTC)

mûre sounds like it comes from Latin morus, morula "mulberry". The mulberry and the blackberry look similar, but are unrelated. The blackberry is an aggregate of drupelets in the rose family; the mulberry is a cluster of individual fruits in the mulberry family, which also includes the fig. The blackberry and raspberry are in the same genus. -phma


cire and wax is not a minimal pair. Someone made I joke, I guess. --129.11.157.69 13:24, 21 Jun 2005 (UTC)

I suspect you're meant to read it vertically, so that cire and sûre are minimal pairs. Could someone add the IPAs? (I don't read French.) Felix the Cassowary 08:40, 22 Jun 2005 (UTC)
Sorry for the mess in history list (pressed some unknown buttton inadvertently). I have added the IPA transcriptions for the French examples. Since IPA is not a fixed pitch font, the layout became quite ugly. So I put the examples in tables instead. Could somebody convert the Hebrew examples to IPA? −Woodstone 20:28, 2005 Jun 22 (UTC)

I used on purpose the colon ":" instead of the true IPA "ː" lengthening symbol because the latter ruins alignment and no confusion is possible. The box containing the symbol is both higher and wider than is normal for the font size, leading to ugly wide horizontal spacing and lifting the line slightly up. −Woodstone 17:42, 2005 Jun 23 (UTC)

Woodstone, not everybody has an identical setup to yours in regards to fonts, OS, rendering system, and browser. We use the correct symbols here to help convince font makers etc of the need for them. In the meantime, the colon is the wrong symbol for people using cut & paste or search. Some solutions would be to fix the fonts used by the IPA template, or set a special font you know works for you for .IPA in your Monobook.css file or equivalent if you are using a different skin. — Hippietrail 05:50, 24 Jun 2005 (UTC)

[edit] Tonal distinction

I deleted the reference to Norwegian and Serbo-Croatian as Tonal_languages to de-emphasize the prominence. Indeed the page on Norwegian mentions tone but mainly prosodic as in other European languages, whereas the article on Serbo-Croatian doesn't mention it at all. On the other side most African languages use tone to a significant extent. Hirzel 14:52, 24 Jun 2005 (UTC)

[edit] IPA for all transcriptions

The transcription of Hebrew and one of the tables for Thai is orthographic rather than in IPA. Please apply IPA to all transcriptions, or it will be hard for those who aren't familiar with the language to understand the differences between the pronunciation. With IPA one will at least have a chance of getting it right.

Peter Isotalo 20:12, 9 July 2005 (UTC)

Done for Thai and Dutch. Who can do Hebrew? −Woodstone 13:20, July 10, 2005 (UTC)

Done. Aheppenh 19:05, 16 February 2006 (UTC)

[edit] Note on Spanish

The comment

In Spanish, [z] and [s] are both allophones of /s/ and [z] appears only before voiced consonants as in mismo /mizmo/.

is unclear for most Spanish speakers from Spain. Although [z] and [s] are allophones in Latin America and Andalucía, they are not in the rest of Spain.

Or am I getting it all wrong? Somebody please, an example in English of [z] and [s] if you think I am getting it wrong.

Nacho --Nachovaca 01:49, 17 August 2005 (UTC)

It's not saying that the sounds spelt z and s (and, for that matter, c before e/i) are the same sound: It's saying that Spanish doesn't distinguish between voiced and unvoiced alveolar fricatives, which are denoted in the IPA as [z] and [s]. A Spanish z (so I'm told) sounds like an English th, written in the IPA as [θ] (if s and z are distinguished). IPA [z] sounds like English S in advise or the Z in English size, zeal. IPA [s] sounds like English C in cell or the S in English size, seal. IPA [c] is a different sound again, but doesn't occur in English or to my knowledge Spanish. IPA sounds in square brackets basically always sound the same, no matter the language they're being used for (at least, they're meant to).
Is that clear? I can be bad at explaining things...
— Felix the Cassowary 04:46, 17 August 2005 (UTC)
Indeed z in most of Spain's spanish sounds like [&theta] but in Latin America and in Andalucia it is an allophone to s. IPA sound should do the trick. The point is the phrase in the beginning of this comment is imprecise.
Nachovaca 19:49, 26 September 2006 (UTC)nachovaca

[edit] Additions on limitations

Many of the recent additions on limitations of minimal pairs seem to be based on a misunderstanding of minimial pairs (e.g. the bit about other information is not relevant to determining the phonemes of a language; if it is regarded as a problem, it must go in Phoneme, because it is a problem of that concept not a tool used to determine the number). Perhaps I merely misunderstand the objection, in which case a reply and an expansion will make me happy.

More significantly, I think they need to be cited. I'm adding an Unreferencedsect template to it.

—Felix the Cassowary (ɑe hɪː jɐ) 14:29, 30 September 2005 (UTC)

[edit] Limitations and flaws of the minimal pair concept and method

The use of minimal pairs to determine phonemes does have its issues. Such a method has the profound weakness of circular logic: phonemes are used to delimit the semantic realm of language (lexical or higher level meaning), but semantic means (minimal pairs of words, such as 'light' vs. 'right' or 'pay' vs. 'bay') are then used to define the phonological realm.

Moreover, if phonemes and minimal pairs were such a precise tool, why would they result in such large variations of the sound inventories of languages (such as anywhere from 40-48 phonemes for counts of English)? Also, it is the case that most words (regardless of homophones like 'right' and 'write', or minimal pairs like 'right' and 'light')differentiate meaning on much more information than a contrast between two sounds.

Still yet another weakness is the very notion of 'minimal'. Minimal pairs of words often, upon closer phonetic analysis, contrast through more than one difference (so the term 'minimal pair' is itself a misnomer). For example, if we break phonemes down further into features, we can differentiate words that are supposed to contrast by one phoneme pair by more than one distinctive feature; however, features have proven even more difficult to delimit than phonemes but some demonstrably spread beyond one segment (which again confounds the very commonly held concept of the phoneme modelling and capturing phonology as a unified sound).

A final possibly fatal flaw of the minimal pair concept is that it is predicated on a psychological phenomemon called 'categorical perception', but categorical perception has been demonstrated as existing as an auditory phenomenon not exclusive to language, which problematizes it as a concept to use in determing a basic phonological unit for language processing and language acquisition.

The above may well be true, but should be formulated in an encyclopedic style, not as a retoric complaint. Perhaps it should also be moved to the Phoneme article. −Woodstone 21:23, 30 September 2005 (UTC)

[edit] Introduction

I've been told that minimal pairs can be made not only between words with distinct meaning, but between one word with a meaning, and a word without meaning. For example, if "abc" has one meaning, and "xbc" has no meaning, then "abc" and "xbc" are minimal pairs, and consequently "a" and "x" are phonemes, even if there are no minimal pairs of "a" and "x" where both words have meaning. Is that correct? Should the introduction of the article be changed to reflect that? Nikola 13:36, 11 November 2005 (UTC)

By definition for something to be a word it has to have a meaning (no matter how vague). There can be a sequence of letters, that is not a word, but there is no meaningful way of establishing its pronunciation, except by following to analogies to words that do exist. So using those non-words to define phonemes would lead to a circular reasoning, which does not lead to valid linguistic results. &minusWoodstone 09:50, 14 November 2005 (UTC)
Would you come to Talk:Montenegrin language#New phonemes and leave your comments, because there is dispute about that? Nikola 19:15, 15 November 2005 (UTC)

Well, then consider the same question asking about [abc] vs [xbc]? Where a,b,c and x are phones used by native speakers. Why do we actually need meaningful utterances? If you insist on meaning i can easily come up with some neologisms to give them meaning. If a native can distinguish the two utterances (no matter the meaning) we have a valid minimal pair, don't we? —Preceding unsigned comment added by 78.130.135.225 (talk) 23:38, 9 June 2008 (UTC)

[edit] Russian language

The Russian language example in 'Differentiating stress' section is bad or even invalid: the vowel 'u' is almost irreducible in normal speech. I would suggest an example with stressed/unstressed 'o', which perfectly and notably reduces to schwa and even to 'a', kind of doroga or some other, simpler word I cannot think of right now. mikka (t) 21:31, 22 November 2005 (UTC)


[edit] Hebrew language

The Hebrew examples in 'Differentiating consonants' are erroneous. These examples are from a reconstructed stage of medieval Hebrew, for which many of the details of pronunciation are uncertain, so they're not a very good choice for illustrating phonetic and phonological ideas. (Sure, they clarify the ideas, even if the facts are dubious, but why not clarify the ideas and at the same time use facts that are reliably accurate?!)

As for the details, */qɔːrɔːʔ/ never existed. This is so for two reasons: First, because at a time when the /ʔ/ was still pronounced (before /ʔ/ in syllable codas was dropped) the vowel /a/ in general had probably not lengthened to /aː/ and /aː/ had not yet shifted to /ɔː/, so this word was /qaraʔ/ or /qaːraʔ/, and second, if the /ʔ/ was still pronounced as a consonant in such a word the vowel before it would be /a/, just as in the other examples, and not /ɔː/. So /qaraʔ/ surely existed at an early stage, and later on /qaːraʔ/, and possibly /qɔːraʔ/, but surely */qɔːrɔːʔ/ never existed. Also, 'to see' was probably /lirʔoːθ/ with a short /i/, so there were two differences in the pronunciation of 'to see' and 'to shoot'. —The preceding unsigned comment was added by 129.49.217.85 (talk) 16:59, 14 December 2006 (UTC).

[edit] excercices

[edit] What minimal pairs can do, and what they can't

English "let" + "lit" proves that phones [ɛ] and [ɪ] do in fact represent distinct phonemes /ɛ/ and /ɪ/.

No, it doesn't. It simply proves that the two do not have the same underlying representation. They could still be allophones of some other phonemes (maybe [ɛ] is the realization of /æ/ before /t/?), or a sequence of phonemes (maybe [ɪ] is a coalescence of /ɛj/?).

If this was true, the IPA chart for English would be missing /ç/ (since "hue" vs. "who" proves it's a phoneme!). --Ptcamn 04:35, 16 January 2006 (UTC)

Same with /ʔ/ in "a nice cold shower" vs. "an ice cold shower". As I write this, I just had the former. ;) --  Denelson83  04:49, 22 July 2006 (UTC)
The ʔ would only appear if you emphasize "ice", and there are more differences. Better ecxample is: "Some mice" or "Some ice", wich is practically the same unless you stress the difference with an unnatural break. Qvasi 10:49, 3 November 2006 (UTC)

I agree completely that lit vs. let does not prove there is a vowel-based phoneme difference. There is proof that the two words contrast, but the location of that contrast in the words is not proved, but is an analytical construct, and what phonologists and phoneticians do is argue about the systm. In this case, they agree on the vowel distinction because it makes for a nice system. "minimal" is the big, big problem in the concept of minimal pair - it is impossible to provide a first minimal pair for an unanalysed language. The issue of "hue" vs. "who" is a case in point. The former word could either starts /hj/ or //ç// - both of which are minimally different from "who" - but in the case of /hj/, where the use of the consonant symbol /j/ implies that the minimality is between /h/ in "who" and /hj/ in "hue", I could analyse the word as containing the diphthong /iu/, in which case the minimal difference is between the vowel /u/ and the vowel /iu/. In addition, some differences are only ever apparently minimal because of the abstract segmental transcription which is used, which hides the many phonetic differences between two words in favour of a single differnece. "Minimal" pair means a single difference in a an analytic description language, not a single difference in sound or articulation, so it must be made quite clear that this is an abstract, theory-led, phonological concept.

[edit] Minimal indistinguishables

What about minimal phone pairs that are indistinguishable by the target language speakers. They prove that the phones in question are part of the repertoire of one and the same phoneme, right? Thus, it's a way of determining the phone repertoire of phonemes. Do you think it's worth including this more explicit wording about determining phoneme constituent phones, in the article? —Preceding unsigned comment added by 78.130.135.225 (talk) 00:01, 10 June 2008 (UTC)