Are there standard words for representing pronunciation?

I have come across a number of questions on the site recently where someone asks how something is pronounced. Someone will give a sample word to illustrate the pronunciation (e.g., a rhyming word for a vowel sound or one in which the relevant consonant sounds the same). Someone else will point out regional variations where the pronunciation is very different, so the sample word is useless as a "universal" example.

An illustration: There was recently a discussion about Mary/merry/marry/Murray pronunciation. I commented about "marry" rhyming with "Barry". Someone pointed out that in some places, "Barry" rhymes with "berry", if I remember correctly. So I offered "Sally" as an example of the "a" sound, just guessing that there isn't much variation in pronunciation.

There was another recent question about "gin" vs. "jean", and tchrist commented, using "fleece" and "kit" as examples. Those seem like pretty safe word choices. Were they good guesses or are those "standard" words that are known to have pretty universal pronunciations?

Most people are not familiar with the letter-like codes used to document pronunciation in places like dictionaries. I assume that there must be a reference to those symbols that says [X] is pronounced like the [Y] in [WXYZ]. The choices for the WXYZ words must be words where at least the relevant portion of the word is universally pronounced the same.

Is there a collection of such "WXYZ" words that can serve as universal examples of what I mean as to how a vowel or consonant sounds?

Answer

"Fleece" and "kit" are part of the Wells "lexical sets." They have been used somewhat as a standard in some discussions about vowel sounds in different dialects, but of course they also have their limitations—Well's use of "palm" doesn't work for American English speakers who have /ɔl/ or /ɔ/ in this word, and the words "cure" and "tour" have different vowels for me and some other speakers (my pronunciation of "cure" sounds to me like it rhymes, or at least almost rhymes, with "fur", but I can't use my "cure" vowel in words like "tour" or "poor").

Wells's lexical sets are based on two artificial standard "reference dialects," "GA/General American" and "RP/Received Pronunciation".

Wells wrote a blog post in 2010 ("lexical sets") where he explains some of the thinking behind his choice of words:

I wanted words that could never be mistaken for other words, no matter what accent you pronounced them in.

Although FLEECE is not the commonest of words, it cannot be mistaken for a word with some other vowel; whereas beat, say, if we had chosen it instead, would have been subject to the drawback that one man’s pronunciation of beat may sound like another’s pronunciation of bait or bit. As far as possible the keywords have been chosen so as to end in a voiceless alveolar or dental consonant…

though that was not always possible.

The least satisfactory keyword is PALM, and its set is also fairly incoherent. Amy says she prefers to replace it with FATHER, which is fine up to a point: but not if we are discussing Hiberno-English, where father often has not the expected aː of Armagh, Karachi, Java etc but the ɔː of THOUGHT.