Is hilarious pronounced /hɪˈlɛriəs/?

Question

For the word hilarious, the pronunciation transcription in the Oxford Advanced American Dictionary is /hɪˈlɛriəs/ but if I click on the speaker icon, I hear /həˈlɛriəs/. Am I listening to it wrong or are Oxford pronunciations not 100% consistent?

Answer 1

The difference between /ɪ/ and /ə/ in unaccented syllables is of very little importance in English. There are some people who use /ə/ instead of /ɪ/ in unaccented syllables (this is called the weak vowel merger). Presumably, the speaker Oxford used had this merger.

More specifically, there are some people who use /ə/ for the first syllables of both hilarious and malarious, and some people who pronounce hilarious with /ɪ/ and malarious with /ə/. Native English speakers typically pay very little attention to the quality of vowels in completely unaccented syllables.

Answer 2

phonemic /ə/ ≠ phonetic [ə]

I realize this may well strike you as an outrageous proposition, but your fundamental mistake lies in somehow believing that dictionaries contain actual phonetic transcriptions. They do not. They contain phonemic transcription, which means that /ə/ does not necessarily mean [ə]. Often it does not.

Here’s where you’ve gone wrong. Dictionaries never contain that phonetic information. They provide only enough phonemic information for native speakers to figure out how to say something in their own accents based on whatever phonological rules apply there.

Per the Wikipedia article on ə:

The symbol ⟨ə⟩ is often used for any unstressed obscure vowel, regardless of its precise quality. For instance, the English vowel transcribed ⟨ə⟩ is a central unrounded vowel that can be close-mid [ɘ], mid [ə] or open-mid [ɜ], depending on the environment.

Notice that it can represent any obscure vowel, one whose precise quality is immaterial.

Phonemically, /ə/ normally represents a set of centralized vowels of indistinct character, particularly when it is reduced and neutralized in unstressed syllables in fluent connected speech. Vowel reduction in unstressed syllables varies significantly by region, accent, register, speaker, and utterance.

Just as the abstract /r/ phoneme can for example represent [ɹ], [ɻ], [ɻʷ], or [ɾ] yet very rarely ever [r], so too does the abstract /ə/ phoneme represent a wide range of possible allophones including [ʌ], [ɵ], [ɘ], [ə], [ɜ], [ɛ], [ɪ], [ɨ], [ɪ̈], [ɨ̞], [i], and probably many more.

So even though these may all well sound hilarious to me but not to you, pronunciations like [hɪˈleɹɪjəs] and [hɨˈlɛɹjəs] and even [əˈlæɾiəs] are all not only the very same word, they also represent perfectly normal allophonic realizations of the very same underlying phonemes.

See also the section covering the weak vowel merger in the Wikipedia article on the phonological history of the English close front vowels.