Allophones – Variation within phonemes
Although English has a limited set of phonemes, there are actually many minor variations in the ways each of these are produced. These variations are not meaningful to English speakers, and so they are often not even noticed by other English speakers. These non-meaningful variants are called allophones. While some variants in the pronunciation of phonemes are caused by idiosyncrasies of individual speakers, or by speakers of different dialects within English, many are determined by the phonological environment of the word in which they occur. As an example, variations are common in the length of time it takes to say certain vowels in English. If we consider the low front vowel /æ/ in English as pronounced in two words, “hat” [hæt] and “had” [hæd], we might notice that the vowel in “had” is held for a slightly longer period of time than the vowel in “hat.” This is a difference in vowel length, or quantity (i.e., there is “more” of the vowel in “had” than in “hat,” so it has a greater “quantity”). We can represent the difference in IPA by adding the [:] symbol to the vowel. So “hat” can be transcribed [hæt] and “had” can be transcribed [hæ:d]. Some languages recognize vowel length as phonemic, but for English the difference between [æ] and [æ:] is not a meaningful difference. They are allophones of the phoneme /æ/.
The reason why we pronounce [æ] longer in “had” is because the word ends in a voiced consonant, allowing the voicing of the vowel to continue into the consonant. In “hat,” which ends with a voiceless consonant, the vowel [æ] must be cut off before pronouncing the [t], causing a shorter length of time in which to pronounce the vowel. The phonological environment of words often affects how we say particular sounds, and while we may not give meaning to these variations, they can often lead to divergences in how the sounds change over time.
Examples of allophonic variation among English phonemes
/p/ (voiceless, bilabial stop)
The words “pit,” “spit,” and “tip” reveal non-meaningful variations in pronouncing the phoneme /p/. When [p] occurs at the beginning of a word, as in “pit,” it is often heavily aspirated (pronounced with a strong puff of air). The allophone can be written [ph], with a superscript [h]. If a word like “spit,” the initial [s] allows us to say the [p] without aspiration. Finally, the [p] in “tip” is often pronounced by closing the two lips but not releasing the air afterwords. This can be represented as [p̚]. In spite of the differences, [ph], [p], and [p̚] are all perceived as instances of /p/ by English speakers.
/t/ (voiceless, dental stop)
The phoneme /t/ has the same three variants as /p/ above, as you can see in the words “top,” “stop,” and “pot,” but it has others as well. Take the words “writer” and “written.” In the first of these, the [t] is medial (in the middle of a word) and occurs after a stressed syllable. In this position in American English the [t] is often produced by a simple alveolar flap, produced by a quick “flap” or “tap” of the tongue against the alveolar ridge. The symbol for this sound is [ɾ]. This sound sounds very much like a [d], so much so that for most people, the words “writer” and “rider” are impossible to distinguish from each other. British speakers, however, preserve the [t] here, so that “writer” and “rider” can be distinguished. Another allophone of /t/ is found in the pronunciation of the word “written.” Many American English speakers do not use an alveolar sound at all for the /t/ in this word, but instead stop the air in the glottis, the space between the vocal cords: this is the glottal stop, written [ʔ]. It is heard in the word “uh-oh,” which in IPA would be written something like [ʔəʔo]. In American English, it appears as an allophone of /t/ whenever the /t/ occurs after a stressed syllable and before /n/ (as in “kitten,” “mitten,” “bitten”). In some dialects of British English, it occurs before or after the approximants /l/ and /ɹ/, as in the words “bottle” [bɔʔəl], “forty” [fɔɹʔi] etc. American English speakers sometimes use it for words that end with /t/. For example, people might say “at” as either [æt̚] with an unreleased /t/ or as [æʔ] with the glottal stop. This makes a total of five common allophones for the phoneme /t/: [th], [t], [t̚], [ɾ], and [ʔ].
/m/ /n/ (labial and alveolar nasals)
The phoneme /m/ is made with two lips closed together, and the alveolar nasal /n/ is made with the tongue against the alveolar ridge. Nasals, however, have a tendency to adapt to the sounds around them. In words like “symphony,” “emphatic,” “infant,” “infinite,” the nasals /m/ or /n/ occur before the labiodental fricative /f/. This sound often influences the nasal sound to also become a labiodental sound, written [ɱ], rather than a bilabial or alveolar sound.
Complementary Distribution
Many of these variations are not random in their appearance. Take the five allophones of /t/ discussed above: [th], [t], [t̚], [ɾ], and [ʔ] (in order, an aspirated /t/, an unaspirated /t/, an unreleased /t/, an alveolar flap, and a glottal stop). The aspirated [th] occurs only in at the beginning of a stressed syllable (top, tin, attend). The alveolar flap only occurs after a stressed syllable and between two vowels, as in writer. The unreleased [t̚] occurs only in word-final environments, like pit. The glottal stop occurs only after a stressed syllable and before [n], as in mitten. All other environments have the “regular” [t]. In other words, in English you will not hear someone saying words like top, tin or attend with an alveolar flap or an unreleased [t̚]; in a word like writer you will not hear someone using an aspirated [th] or an unreleased [t̚]. In a word like pit, you will not hear someone using an alveolar flap or an aspirated [th]. In other words, these variations occur in complementary distribution: one sound never occurs in the environments in which the other one occurs. In fact, complementary distribution underlies the very essence of allophones. Because two sounds do not occur in the same phonetic environment, they will never contrast with each other; they will never occur in contrastive distribution. If you remember from last week, two sounds that meaningfully contrast with each other, such as in a minimal pair, are called phonemes. (Note: do not confuse complement, from “to complete” with an “e” and compliment, meaning “to praise someone” with an “i”).
To give another example, consider the two words tip and pip: [thɪp̚] and [phɪp̚]. The initial [th] and initial [ph] occur in the same phonetic environment—the beginning of a one syllable word (word initial position). They contrast with each other so that we can meaningfully distinguish them: One is /t/ and the other is /p/. Because we recognize the two words as having different meanings, we know that these two words form minimal pairs and thus we know that /t/ and /p/ are different phonemes in English. Now look only at the word pip: [phɪp̚]. The first sound [ph] is a sound that only occurs in word initial position. The last sound [p̚] is a sound that only occurs in word final position. (The pronunciation *[p̚ɪph] cannot occur in English). There will never be a minimal pair in which [ph] will contrast with [p̚]; instead, the two sounds complement each other. In other words, sounds that occur in contrastive distribution are phonemes and sounds that occur in complementary distribution are allophones.
One reason this is important for this class is that when sounds change over time, it is often only a particular allophone that changes. As an example, let’s take two allophones of the phoneme /k/, as heard in the words “leak” and “lock.” In the first one, “leak,” the /k/ follows an /i/, which is a high, front vowel, but in the second, the /k/ follows /ɑ/, the low, back vowel. These vowels, produced in such different places in the mouth, will have an effect on how the /k/ is said, producing two allophones. In “leak,” as you might expect, the front vowel /i/ “pulls” the velar /k/ towards the front of the mouth, so that it is really said on the hard palate rather than the velum. We could mark this type of /k/ as [k’]. On the other hand, the low, back vowel in “lock” keeps the /k/ far back down on the velum. We will write this one as [k]. In Present-Day English, this has so far had no effect on the phoneme /k/. If we go back in time 1500 years, to early Old English, we will see that it had these two allophones of /k/ too, for the same reason. The “fronted” [k’] occurred when it was near a front vowel like /i/ or /e/, and the regular, velar [k] occurred when it was around low or back vowels. Today, most of the words we have with a “ch” sound, /tʃ/ come from the fronted [k’] in Old English: “church,” “chin,” “cheap,” “choose,” “rich,” “teach.” And many sounds we have with /k/ (“back,” “lock,” “leek,” “cold,” “cool”) come from Old English words in which the /k/ was not near a front vowel.
If we had to guess about the future, we might think the same thing will happen with the alveolar flap, as in the word “writer.” We perceive the medial sound in this word as /t/ for several reasons. First, as English speakers we know that the base word on which “writer” is built is “write,” which has a very clear /t/ (even if it is an unreleased [t̚], or even in some pronunciations a [ʔ]). As literate speakers, we also know that we spell the word “writer” with the letter , which influences how we perceive the sound (Note that letters of the alphabet are often written between angled brackets < >). But if we listen only to the sound of the word “writer,” as perhaps a foreign speaker might, we would realize that the sound is very close to [d], as in the word “rider.” We could formulate a rule from this that all instances of /t/ that occur after an accented syllable and between two vowels merge with /d/, as in the words “writer,” “lighter,” “latter,” “fatter,” “matter,” “motor,” “meeting,” etc. Each of these are pronounced with the alveolar flap [ɾ] that sounds like [d]. To test this, create minimal pairs. Can you tell the difference between writer – rider, matter – madder, slighter – slider, beating – beading. For most American English speakers, the answer is probably no (which technically means that the alveolar flap in these sounds is not a different phoneme from /d/). But have all instances of /t/ become /d/ in American English? No, only those instances where we are dealing with the alveolar flap.
