While the phoneme is the abstract unit or sound-type (“in the mind”), there are many different versions of that sound-type regularly produced in actual speech (“in the mouth”). We can describe those different versions as phones. Phones are phonetic units and appear in square brackets. When we have a set of phones, all of which are versions of one phoneme, we add the prefix “allo-” (= one of a closely related set) and refer to them as allophones of that phoneme.

For example, the [t] sound in the word tar is normally pronounced with a stronger puff of air than is present in the [t] sound in the word star. If you put the back of your hand in front of your mouth as you say tar, then star, you should be able to feel some physical evidence of aspiration (the puff of air) accompanying the [t] sound at the beginning of tar (but not in star). This aspirated version is represented more precisely The sound patterns of language 43 as [tʰ]. That’s one phone.

we noted that the [t] sound between vowels in a word like writer often becomes a flap, which we can represent as [D]. That’s another phone. In the pronunciation of a word like eighth (/etθ/), the influence of the final dental [θ] sound causes a dental articulation of the [t] sound. This can be represented more precisely as [t̪]. That’s yet another phone. There are even more variations of this sound which, like [tʰ], [D] and [t̪], can be represented in a more precise way in a detailed, or narrow, phonetic transcription. Because these variations are all part of one set of phones, they are referred to as allophones of the phoneme /t/. The crucial distinction between phonemes and allophones is that substituting one phoneme for another will result in a word with a different meaning (as well as a different pronunciation), but substituting allophones only results in a different (and perhaps unusual) pronunciation of the same word.

Let’s look at another quick example, using a vowel sound. In English, there is a subtle difference in the pronunciation of /i/ in the words seed and seen. In the second word, the effect of the nasal consonant [n] makes the [i] sound nasalized. We can represent this nasalization with a small mark (˜), called “tilde,” over the symbol [ı̃] in a narrow phonetic transcription. So, there are at least two phones, [i] and [ı̃], used to realize the single phoneme. They are both allophones of /i/ in English. It is possible, of course, for two languages to have the same pair of phonetic segments, but to treat them differently. In English, the effect of nasalization on a vowel is treated as allophonic variation because the nasalized version is not meaningfully contrastive. Whether we try to say [sin] or [sı̃n], people will only recognize one word seen. In French, however, the pronunciation [so] for the word seau (“pail”) contrasts with [sõ] for the word son (“sound”) and beau [bo] (“good-looking”) contrasts with bon [bõ] (“good”). Clearly, in these cases, the distinction is phonemic.