Phonemes and allophones

This video just has a plain transcript, not time-aligned to the videoThis video introduces the notion of phoneme as a basic unit of phonological analysis. While phoneticians are interested in describing and modeling the physical details of speech from a variety of perspectives, phonologists are interested in patterns of sounds. In particular, phonology addresses patterns and distributions of phones within individual languages as well as across languages.
A basic concept of phonology is that we need (at least) two levels of representation to adequately represent and describe the patterns that we observe. One of these levels is called the surface or allophonic level represents something close to articulation and the phonetic descriptions that we’ve been learning so far. The other level is called the underlying or phonemic level, which represents abstract categories that are something like our perceptual judgments about which sounds are and are not similar to each other. Somewhat confusingly, both of these levels use symbols from the IPA. In order to distinguish between the two levels of representation, we use two types of brackets. For the surface forms, we use square brackets, and for the underlying forms, we use slashes. It is important to note here, that although [ ] can indicate varying degrees of detail, / / can only indicate abstract categories of phonemic contrast.
Now let’s consider why we need these two levels of analysis. Take these two words of English: met and net. First, let’s transcribe their surface forms. Now that we have a representation of how these words sound, we can see that on the surface these words only differ in one phone, at the start of the word. Other than place of articulation, [m] and [n] are phonetically very similar to one another. However, they sound categorically different to English speakers. In fact, they are so different to English speakers that this small change from one place of articulation to another is enough to single a difference in meaning between these two words. In other words [m] contrasts with [n].
In fact, we can even put [n] and [m] into novel strings of phones (that is, unfamiliar words or even non-words) and English speakers will still hear the difference between these two sounds.
Let’s consider another example. Take the words ten and tenth. Again the first thing we need to do is transcribe these words phonetically, representing their surface forms. In this case, what we see is a difference between an alveolar nasal and a dental nasal. Once again, these phones are phonetically very similar, differing only in place of articulation. However, in this case, these phones are categorically the same to English speakers. In fact, most speakers probably don’t even notice that they are using different articulations for each of these sounds. So how are we to understand why a change in place of articulation is enough to ”sound different” to English speakers, while another change in place of articulation sounds the same?
One answer you might want to give is that [n] and [n̪] are just too similar. However, there are languages that use these phones to distinguish between words in just the same way that English uses [m] and [n]. For example, in Mapudungun, an indigenous language of Chile, the subtle change from alveolar to dental place can signal a drastic difference in meanings as we see in the word pair here. So we can’t say that this difference is just too subtle because there are languages that use it to signal changes in meaning. It just happens not to be a relevant difference to English speakers.
So, why can’t English speakers hear the difference between [n] and [n̪]? In order to explain this, we need to introduce a second, more abstract level of analysis, the phoneme. This abstract representation of a nasal, symbolized as the letter n between two slashes, indicates a group or category of sounds that belong together in the mind of speakers, but have different surface representations. These surface representations, represented between square brackets are known as allophones and they are language specific.
These surface representations, represented between square brackets are known as allophones and the relationship between phonemes and allophones are language specific.
We can represent this relationship in phoneme diagrams such as the ones shown here. These demonstrate that English has one phoneme with (at least) two allophones, while Mapudungun has two separate nasal phonemes.
When a phoneme has more than one allophone, those allophones DO NOT contrast with one another.
Now that we understand the basic concepts of phonemes and allophones, we need to be able to describe the distributions of allophones, as well as how phones are organized into phonemes.
In order to answer these questions, let’s look at some words containing these phones.
Take a moment to consider the words transcribed at left. (pause the video if you need more time). As you do think about what the pattern is in distribution of these two phones. When we're looking at the pattern of distribution we're thinking in particular about the phones that surround the sounds that we're interested in and also consider whether we can predict which phone will occur how can you predict it in what way can you describe the predictive factor.
Pause the video if you need more time.
Phonologists sometimes formalize this relationship between the phoneme and its allophones in a rule such as this one. The notation here means that the phoneme /n/ is realized as a dental nasal before either θ or ð, and as an alveolar nasal anywhere else
The arrow is read as “is realized as” and the slash stands for “in the environment of”. The blank shows where the phoneme occurs in order for the rule to apply. In order to fully define a phoneme, we first need to observe the surface forms that occur, along with their environments. Then, we need to describe the patterns that we see with respect to the surface forms and their phonetic environments, looking for generalizations along the way. The types of generalizations that we typically mean are those having to do with shared features across the predictive environments. For example, what is the feature that is shared among θ, ð, and n̪?
Pause the video if you need more time to consider.
A clue to the answer here is in the names of the phones themselves: the dental nasal occurs before a voiceless dental fricative or a voiced dental fricative. So we can actually make this rule more general by referring to these shared features
Now consider this further data. Can you write a rule to describe which surface form of /n/ will appear? One way to start here is to simply write out a rule for each item in the lexicon. So here I have identified the nasal in each word and written it in formalized rule notation indicating the environment in which it occurs. In this case I’m only considering the following environment although sometimes the preceding environment can be predictive as well.
Once we've done that we can think about what features the surface forms share with their predictive environments. So for example in the first word sank the velar nasal occurs before a velar stop. Both of these phones share a place of articulation: velar.
If we go through this process in its entirety we can see that the /n/ phoneme takes on the place characteristic of the consonant that follows it.
We won’t go into the formalisms necessary to write that more abstract rule here, but we can still state the generalization in prose like this. The phoneme /n/ (between slashes) is realized with the place of articulation of the following consonant. This is a kind of rule known as assimilation and we'll learn more about that in future video.
Now that you have an idea of how to proceed, take some time to think about another phoneme in English, this time /t/. In order to define the phoneme we need to think about what its surface forms are, what environments they occur in, and whether we can predict where these surface forms occur can you write a rule what is it. How many rules do you need to describe the allophones of /t/? This probably depends on the number of surface forms and the environments that you occur in. What kinds of factors might influence your answers to the above? Is there only one answer, or are there many?