 The central question in phonology concerns the basic elements of sound structure. A very popular view in this respect is that it is the phoneme, that is the central element in phonology. To understand what a phoneme is and what it is not, we will proceed as follows. We will first of all perform an analysis of an input string. Then we will ask about the function of each particular element in that string. We will then perform a phonetic, or to be more precise, an articulatory analysis of that string and the elements within it. And we'll finally look at the distribution of some selected phones or sound units within that string. So let's start with our string. Here it is. Liz played with Paul. And now I want to be as precise as possible in transcribing the phonetic details contained. In other words, I have to apply a narrow phonetic transcription. And to signal such a transcription, we first of all need these angular brackets. And I will first use the set of symbols as supplied by the International Phonetic Association. And then in the next step, I will add diacritic elements to add additional information. So here are the symbols. Liz played Paul. Now, in a narrow phonetic transcription, I have to add diacritics to mark all phonetic details that are relevant. For example, we know that the alveolar lateral consonant in Liz is slightly palatalized. Liz, the diacritic for palatalization, is a superscript J. Then the le in played, the alveolar lateral in played is also palatalized. But at the same time, it is slightly devoiced. If you pronounce that very carefully, it is played. Then the other elements are pretty clear. But here we have a typical example of a plosive with full aspiration. So we don't say Paul, but Paul. And this little puff of air that accompanies a plosive at the beginning of a word in English, a voiceless plosive, can be marked by the superscript H, which stands for aspiration. And finally, the last element, this alveolar lateral, again, is now different. It's not palatalized anymore, but it is velarized, which you can mark by this additional diacritic. So this is the first result of our analysis. What are we going to do with this analysis? Well, first of all, let's look how these elements function within particular words. So here I have our velarized alveolar lateral, and I want to contrast it with some other elements. So let's first of all look at, here we have Paul in the first case with our velarized element, Paul. And now if I pronounce it Paul, again, we would understand the name Paul. Only people would tell me, well, what sort of pronunciation is that if you say Paul instead of Paul? You sound like a foreigner, don't you? But people would clearly understand the meaning Paul. So it doesn't really matter whether we use the palatalized one or the velarized one, the meaning will be Paul. But what about Paul versus Paulk? In American English this would be pork, but we're talking about British English here. So Paul versus Paulk in terms of transcription, this would be Paulk and this would be our Paul. So here clearly these two phones result in a differentiation of meaning. And the same applies to the next case if we replace the final element by z. So here we have our Paul again, and now we get Paul's. This will be the word. So in other words, here we have again sound units that result in a distinction of meaning if you put them in the same position. In other words, we have sound units that must be different somehow. We have the k on the one hand, we have the z on the one hand, and we have our l. Since we get differences in meaning in the function, they constitute different so-called phonemes. That's the first approach, and by the way these units that differ in one position are called minimal pairs. But what are we going to do with our alveolar laterals? Well let's analyze them acoustically in more detail. They are all voiced, normally voiced alveolar lateral consonants. Well this one is d-voice, but that's slightly different. So let's look at the articulation in more detail. So here we have the first one, and let's draw the exact tongue position which somehow looks like that. So it's palatalized. So here in the palatal region we have some sort of narrowing. This one is palatalized too, but it is d-voice. So let's draw this circle underneath which tells us this must be d-voiced. Well and this one, so that was the one in Liz, that was the one in Play, and that was the one in Paul. Well and this one looks like this here. So this is the tongue configuration. Now if we move our face back here, we can see they are phonetically relatively similar. They are produced with a similar articulatory posture. Only the back of the tongue is raised here in the velarized alveolar lateral. The tongue positions are identical in these two, but this one is d-voiced. So in other words our three phones are all phonetically similar. This concept is called phonetic similarity. Let's now look at the position of each of our L's. Where do they occur? Well the first alveolar lateral occurs before vowels. This underscore, an underscore in linguistics is normally stands for the position of the element that you have in focus. So this would be our palatalized lateral. And the next element must be a vowel. So V stands for vowel. Liz, like, look, lose and all sorts of vowels can occur here. The alveolar lateral that precedes must be palatalized. Let's look at this one by contrast. This one occurs after vowels as in Paul. So it is this position, positional effect. As in Paul or cell or deal. So all sorts of vowels can precede. If you have an alveolar lateral that follows it must be velarized. Well this one here also occurs, is also palatalized, but it occurs after two consonants. Namely after P and C as in play or clay or claim or clear or, well, plumage. So in other words we have very precisely definable contexts. These contexts are complementary. The positions are complementary. One of these alveolar laterals occurs, we wouldn't expect another one. If we swap them, the result would just be a very strange pronunciation. So something like Paul instead of Paul or something like played instead of played. And something like Liz instead of Liz would just sound very awkward, but the meaning would not be affected. So they are phonetically similar and they occur in what you call in complementary distribution. Well they behave just like two famous characters that we know from literature. In the novella by Robert Louis Stevenson, the Scottish author, Dr. Jekyll and Mr. Hyde we have one person with a sort of split appearance. Dr. Jekyll only occurs at night time and Dr. Hyde only occurs at daytime. In other words these two appearances are in complementary distribution. Where we have one of them we wouldn't expect the other. So what are we going to do now with these findings? Well we can now define these three phones as the members of one family. They occur in complementary distribution. They are phonetically similar so they constitute one family and this family is called the phoneme. Now the phoneme is represented in these square brackets. And now we have the choice. What sort of symbol can we take? The phoneme is of course an abstract unit. We do not pronounce phonemes, we pronounce phones. And the abstract unit is normally represented in terms of the most simple character that we have available in the Roman alphabet. So here we would suggest the L without any diacritics as the symbol for the phoneme. This view is called the physical view of the phoneme and it was first defined by this man, Daniel Jones. Many people would say the father of English phonology at the beginning of the 20th century. Now with phonemes we can now reanalyze our initial string. So when we now write down the string in terms of a phonemic, a broad phonemic transcription, we can simply do the following. Liz played, oops, so that one is wrong, played with Paul. Now we are using the phonemic brackets. Liz played with Paul. And now we know of course that these phonemes here have to be realized by particular alophones depending on where they occur. Here we have the context before vowels, it must be I. Here we have the context after voiceless plosives, it must be. And here we have the context after vowels so it must be the all variant. Well, this is the phoneme. An abstract head term with its family members can be one, two, three or several, even more in some languages. Its members are phonetically similar. So we have the concept of phonetic similarity and they occur in complementary distribution. The members of the family are called phones and since they belong to the same family, they are not just called phones but alophones. So here we have one basic element of sound structure in phonology, the phoneme in its physical interpretation.