The problem with "the dictionary" as arbiter

I missed the Weekend Edition Sunday Puzzle last week, but (as usual) the answer to last week’s challenge was broadcast this week. Here‘s what the challenge was:

In most words containing the letter “O” between two consonants, the O is either pronounced as a long O or a short O. Can you name a common word in which O appears between two consonants and the O is pronounced like a short “i”?

The answer is not surprising, but I’m a little annoyed by the reasoning used to justify the claim that the answer is unique.

Answer: Women (Words such as cinnamon, pivot, parrot, ribbon, common, havoc and cotton are what the dictionary calls “schwas” or words with an unstressed vowel sound. Therefore the only correct answer is “women.”)

I don’t even know where to begin here. First, there’s the unspecified “the dictionary”. Then, “schwas” are defined as “words with an unstressed vowel sound” rather than as the unstressed vowel sounds themselves! (To be fair: Will Shortz didn’t make this latter mistake on the air, as you can hear here.)

But of course, the main problem is that schwa ([ə]) is not the only unstressed vowel sound in English, dictionary arbitrations notwithstanding, and in fact for many speakers the “short ‘i’ sound” ([ɪ]) is commonly used in contexts where other speakers might have [ə]. For example, I pronounce the words listed above as follows (ń = syllabic nasal, 1 = primary stress, 2 = nonprimary stress, 3 = unstressed):

word pronunciation stress pattern
women [wɪmɪn] 1-3
cinnamon [sɪnəmɪn] 1-3-2
pivot [pɪvɪʔt] 1-3
parrot [pæɹəʔt] 1-3
ribbon [ɹɪbɪn] 1-3
havoc [hævəʔk] 1-3
cotton [kɔ:ʔń] 1-3

FWIW: in words like parrot and havoc, I prefer [ə] but allow [ɪ] as an alternative. By contrast, in none of the cases where I transcribe [ɪ] do I allow [ə].

There are probably patterns to my pronunciation of [ə] vs. [ɪ] in unstressed positions; in other words, I don’t think I have two phonemically distinct unstressed vowels in my inventory. My point is that those folks who submitted these “incorrect” answers were using their noggins ([nɔ:gɪnz] 1-3, not [nɔ:gənz] or [nɔ:gńz]) rather than their dictionaries, and they were right to do so given the statement of the challenge. In order for “women” to be the only correct answer, the challenge should have been stated either like this …

Can you name a common word in which O appears between two consonants and the O is pronounced like a stressed short “i”?

… or like this (putting aside the unspecified “the dictionary” bit):

Can you name a common word in which O appears between two consonants and the dictionary says that the O is pronounced like a short “i”?

2 thoughts on “The problem with "the dictionary" as arbiter

  1. John McCarthy

    In his book English Sound Structure, John Harris says that there are “two main types of vowel-reduction systems in English”, which he designates as A and B (p. 110). In system A, all vowels reduce to schwa. In System B, some vowels reduce to [ɪ]. These are the vowels that alternate with stressed [i:], [ay], and [ɛ]: demon/demonic, horizon/horizontal, telepathic/telepathy.

    As is his practice throughout the book, Harris doesn’t identify these dialects beyond the enigmatic A and B. I would imagine that someone has done extensive empirical work on this topic and has published it somewhere, but I haven’t seen it.

  2. Jarek Weckwerth

    Funny, John Wells writes about unstressed vowels in his Sept 12 entry, saying that southern hemisphere Englishes and most rhotic accents have a preference for schwa (rather than [ɪ]) in unstressed syllables. He also says that there has been a similar trend in England, making e.g. “chartered” and “charted”, or “affect” and “effect” homophonous. I seem to remember that he discusses this in more detail in his Accents of English but I don’t have a copy at hand at the moment. (And the book isn’t particularly recent…)

Leave a Reply

Your email address will not be published.