Phonemes are often defined as the sound units that make up spoken words — specifically as the smallest unit of speech that serves to distinguish one utterance from another (e.g., BAG and RAG are distinguished by their first phonemes). However, this is not completely accurate. In actuality, speech does not consist of a series of discrete sounds. For example, we say that BAT has three phonemes (and SPAT has 4) because it is a useful level of description. But if you record the word BAT you will find that you cannot break the recording up into three discrete parts for the three phonemes. Instead, the sounds of “b” “a” and “t” blend together (for more, watch the second video on this page: the lost phonemes of bat). As another illustration of this blending, try saying the word SOAP and pay attention to the shape of your lips before you even say the word. Now start to say the word SEEP and notice the shape of your lips. Contrast the two. Before you even say the two words your lips are in different positions because of the different upcoming vowels.
So if speech does not actually consist of a series of discrete sounds, what are phonemes? “Phonemes” are an abstraction and a simplification. Phonemes are the units that make up an abstract code that develops in the brain as a person learns to link alphabetic writing to the spoken language they already know. Alphabetic writing differs from speech in a very basic but important way. Unlike speech, the spellings of words do consist of a series of discrete units – the letters. Learning how these two different ways of representing language are related is an interesting problem, and an early obstacle for some beginning readers. Phonemes are the solution to this problem. By learning to treat speech as if it is made up of phonemes, beginning readers can learn how to represent spoken language with an alphabet. For more on phonemes check out this video on phonemes, speech, and reading.
“Phoneme” being an abstraction, it works very well most of the time but not always.
For example, we can agree how many phonemes are in simple words like “bat” “cheat” “token” and “philosophy”. (3, 3, 5, 8: ph/i/l/o/s/o/ph/y).
But there are many cases that are very unclear. How many phonemes are there in “fir”?
It is usually treated by linguists as having 3 phonemes, using a test like this: “fir” and “sir” differ by one sound. therefore, the initial “f” is a phoneme. “fir” and “far” differ by one sound.
Therefore, the vowel is a second phoneme. “fir” and “fin” differ by one sound. therefore the “r” is the third phoneme.
But you can see that the last example is not the same as the first two – the vowels are also different in “fir” and “fin”. When teaching children, “fir” is usually treated as having 2 phonemes (/f/ and /er/) while “fin” has 3 (/f/i/n/). The fact that their spellings are similar—both are CVC, consonant-vowel-consonant syllables—might lead you (or a child) to think they have the same number of phonemes, but spelling can be misleading.
Here is another unclear case: “fail”
Is it three, like in “feel”?
Linguists will usually say four because they treat the vowel in “fail” (a diphthong, a type of vowel) as two phonemes. However, when teaching a child to read, the sound will probably be taught as the “Long A”.
These weirdnesses are NOT your problem! They just illustrate that treating speech as consisting of phonemes is a simplification of actual speech.