 One of the most common complications of simple morphological analysis is that morphemes may have different shapes and different circumstances. This phenomenon is referred to as morphological alternation. The subfield within morphology is called allomorphine. Languages differ enormously with regard to allomorphine. For this reason, this unit exemplifies the phenomena of morphological alternation, not only on the basis of English, but in a cross-linguistic fashion. Consider the following example, the indefinite article in English. We have three variants. We have the variant, which you might want to transcribe as a. We have the variant a, and we have this variant. Now, you see I have to represent this phonologically. Now, the first variant can be found in examples like a book. This one can be found in particular contexts where you say something like, I saw a fantastic concert. So, this is where you emphasize the article, and this one can occur in contexts like an idea. In other words, we have three allomorphs that represent the same morpheme, which we might want to write down like this. How do we know when to use which allomorph? Well, here we can clearly find conditions which I'm not going into now. Depending on the type of allomorphy, various conditions for the occurrence of particular allomorphs may be defined. So, look at them in more detail. We have, for example, phonological allomorphy. That is, the alternation can be described in terms of a phonological rule. Now, our example were, of course, our three variants where we know that this one here occurs before vowels like an idea, an answer. This one occurs in stress environments like a concert, and this one occurs elsewhere. So, here we have a phonological condition. Weak, supportive allomorphy is a kind of morphological alternation where the allomorphs exhibit some phonological similarity, but their relationship cannot be described by a phonological rule. So, in English, we have examples like by and bought. So, phonologically, you have this I becoming or by bought. But that's, of course, nothing completely regular, so we cannot set up a phonological rule or alternatively in meet and met. That's impossible. However, there's some sort of similarity, isn't it? Even less similar are the allomorphs of a particular morpheme if we have the case of strong, sublative allomorphy. So, here examples would be go and went or cases like B is R and so on. Here we have morphs belonging to the same paradigm, but they do not exhibit any similarity at all. You call this type of allomorphy strong, sublative allomorphy. Now, in addition to these main types, we can now define various conditions under which a particular allomorph is selected. For example, phonological conditioning where the sound structure conditions the choice of allomorph. For example, in English we have things like, and now I'm using transcription, horses versus cats. So, the choice of the pluromorph depends on a phonological rule. A case of morphological conditioning can be found if the morphological context determines the choice of a particular allomorph. So, for example, if you look at the, and now again I'm using phonetic transcription, wife. The genitive requires wife, so let's write down genitive here. Whereas in the plural, you have wives, so this is the plural form. The genitive is clearly phonologically conditioned, whereas the plural form is a special morphological context which determines the choice of the allomorph. And finally, we have lexical conditioning. Now here, again, we can use our example by and bought, where the choice of an allomorph clearly depends on a particular lexeme. So, in by and bought, we have this sort of vocaliic relationship, but if we contrast that with cry, it is of course cried. And not as we might want to have assumed crot, which is of course not the past tense form of cry. So, here we have lexical conditioning where the choice depends on a particular lexeme. Let's now look at these cases in more detail. Phonological conditioning is the first of these. Consider the following word pairs, so let's use cat and cats. I'm now using orthography, dog and dogs, and horse, horses. So, we have plural, affixes, sir, sir, and is. Now, in all three words, the plural morph is phonologically different. However, the distribution of these allomorphs is complementary and can be predicted on the basis of the final stem consonant. So, in cats, we have a t that precedes. In dog, we have a g and in horses, we have a sir. Now, if you look at these in more detail, you will find that the t, this one is voiceless. The g is voiced plus voice, and the sir is a special case. It is an alveolar fricative consonant, and this consonant, a voiceless fricative consonant, requires that is is added. Observations of this kind can be generalized and can be converted into morphological rules. Let us look at some other languages now. In Hungarian, the morphine that denotes the locative relationship in, like in the book, has two allomorphs, ben and barn. Here they are. So, in the wall versus in the air. The occurrence of these two allomorphs is clearly phonologically conditioned. Ben occurs if the final stem vowel is a front vowel, barn has to be chosen if the final stem vowel is a back vowel. In Turkish, we have a very similar situation. Now again, the choice depends on the type of stem. We have gözler, in which case we have a front vowel and ler, and we have dudaklar, which is a back vowel and lar. This phenomenon that we have just seen in Hungarian and in Turkish is a phenomenon where the vowel of an allomorph has to be harmonized with a stem vowel. And this phenomenon has become known as vowel harmony. Let us look at morphological conditioning next. Morphological conditioning occurs if the choice of allomorphs is determined by the morphological context. Here is an example from Latin. We have the base form loud, which means to praise in laudare, which is the infinitive. Now, the first person singular present tense is laudo. The first person in the simple past is laudabam. And the first person in the present perfect is this one. In other words, we have three morphs that denote the first person singular. Now, the choice of these three forms is of course dependent on the morphological context. If we have just the base form that precedes in the present tense, we have just o as the marker for the first person singular. If the past tense occurs, then m has to be the realization of the first person singular. And if we have the present perfect variant, then we have to add e. The choice of a particular allomorph is dependent on the morphological context. In English, this can be typically found in forms such as wife, where we either have wives, the genitive, or wives, which is the plural form. So here, the choice of a particular form depends again on the morphological context. Let's finally look at lexical conditioning. Now, cases of lexical conditioning occur where the choice of an affix is dependent on other properties of the base, for example, on semantic properties. Here is an example from Persian, where plural marking depends on the semantic properties, human versus non-human, of the base. So we have a form, and now I'm using alphabetical transcript, mer'd, which means man, and we have gorbe, which means cat. Now, if we want to turn these into their plural forms, we have to add an mer'dan to mer'd to man, and we have to add ha, which is also the plural, to the base form of cat, gorbe ha. The choice here, or the determining factor, is human plus human and minus human in cat. Lexical conditioning is also defined for cases where the choice of alomorphs cannot be derived by any rule at all, and has to be learned. This applies to the English past participle affix en, where speakers have to learn which verbs take this suffix, taken, seen, given, and so on. So let's summarize. Several conditions can be postulated that allow us to determine the choice of alomorphs and to set up morphological rules. The discovery of such rules, especially phonological rules, so rules of this type here, phonological conditioning, the discovery of such rules is one of the major goals in the morphological analysis of the languages of the world.