 Language is a communication system employing arbitrary symbols. These symbols, normally words, have to be stored. The following different types of word store exist. On the one hand, we have dictionaries. On the other, we have lexicons. Well, both types of word stores will be addressed in the following. Let's start with dictionaries. Now, word stores that are primarily consulted for the reason of retrieving information about the words of a language are referred to as dictionaries. Today, we have two types of variants of dictionaries. We have book dictionaries over here, the classical form of a dictionary. And we have machine readable dictionaries. Dictionaries which are delivered either on external disks such as CD-ROM, DVD, or more commonly today on the web. The most common types of book dictionary are encyclopedic monolingual dictionaries, bilingual dictionaries or dictionaries for special purposes, such as dictionaries that represent synonyms. Machine readable dictionaries are similar concerning their content. However, as I said, they are delivered externally on specific storage devices such as CD-ROMs or in the cloud of the World Wide Web. Dictionaries do not play any role within the theory of grammar. Rather, they are independent word stores used to obtain information about the words of a language. Let's now look at lexicons. Now, a lexicon is the central component of a natural language processing system, whether human or machine. And thus, it is a central component of a theory of grammar. Why is this so? Well, any theory of grammar seeks to define among others the principles underlying language processing and language acquisition. Thus, the central word store in this respect is the mental lexicon rather than any variant of a dictionary. The reason is quite simple. Whereas dictionaries are static and do not allow any changes to their contents. The mental lexicon, the lexicon we have in our minds or the lexicon that is used in a grammatical theory, permits the possibility of changing its structure, additions, extractions, rearrangement of the material or the modification of the entries. It closely interacts with other components of grammar and provides detailed information about the words to be generated. Let's see how this information, the information associated with lexical entries, is represented. And we will look at all levels of linguistics in the following from phonology to semantics. So let's start with phonology. Now here I have four lexemes built C and that and I would now like to build up the phonological information associated with each entry. The phonological specification in the lexicon defines a segmental and suprasegmental properties of a lexeme. That is, we have to present the phonemic structure for example. So here the fact that we have bill consisting of these phonemes. We could also represent allophonic information or distinctive feature information, but let's keep it as simple as possible. And we have to represent information about the syllable structure. So here in this case it is a typical syllable with an onset, a peak and a coda. We might want to add stress aspects, intonational aspects. Well here we don't have to do it because we have a monosyllabic entry in present day English. Let's do the same for the other entries. Here we have C which is this sort of syllable structure. And we have that which, well that is an interesting element. It can be a weak form so in connected speech this behaves a little bit different from the others. And finally we have the preposition in which has just a V-C structure. So this then is the phonological representation of our four lexemes. Let's now look at the morphological information associated with our lexemes. Now the morphological specification defines the lexical category. That is the word class associated with each lexeme. Now this would be very simple. However we need a little bit more information than that bill is a noun. For example we need a special noun class. Let's call it noun class 2 which indicates to us N2 would be an arbitrary symbol by the way. Which indicates to us that under normal circumstances bill does not involve a plural. Likewise we have to represent that C is a verb but again we could use a symbol let's call it V4. Again this is fully arbitrary we could call it V12345 depending on our internal classification. Now here we want to indicate that verbs of this particular type have an irregular past form that has to be retrieved from elsewhere. So that is a determiner and N is a preposition. So this is the morphological information. Let's now move on with syntax. The syntactic specification defines the context of a lexeme in terms of other categories. The notation is relatively straightforward. Now first of all we define the word class. Bill is a noun and here we don't have to represent the fact that bill is a special noun that doesn't involve a plural under normal circumstances. We simply write down N for noun. Then a colon follows this is a notational convention and then we have the context in angular brackets. Now what is the context of the noun? Well this is the position of the noun itself. This is where the noun occurs. Can anything occur after the noun? Bill by brother? Bill from London? Well it can almost unrestrictedly. However before the noun we have to represent the fact somehow that a determiner cannot precede. So the bill is normally impossible. Well of course you can have sentences like the bill I know is much stronger than the bill you know but that's a special context. Let's now look at the verb C. Well C is a verb and well this would be the position of the verb and underscore for the position. Well and what can follow? Can I have a sentence like John C's Mary saw? Well I can't. What we do need is a noun phrase that follows. Well and this frame this representational format is called a subcategorization frame. Likewise we can represent this information for that which is our determiner. Well what is the context of that? Well that has to be followed by a noun that is represented in the singular. If we take this particular that as a determiner and not as a conjunction. Well and finally we have our preposition. Now in English prepositions as the term says precede something. Now what do they precede? Well they precede obviously a noun phrase. Well and this noun phrase well it has to be locative or directional. So in the garden, in London and in five minutes which is of course temporal but time information can also be interpreted as a special case of location locative information on a time scale. So this concludes the syntactic representation or the syntactic specification of a lexical entry. Let's now move on with the meaning that is the semantic specification. Well this specification defines the meaning of a lexeme. This can be done via lexical features in a binary fashion plus minus values or it can be done via special symbols that are coupled with more complex information. Well let's do it for our examples. Well bill first of all is of course our noun again and now we can of course associate some features with this. So bill is animate, bill is human, bill is male. Well let's stop here. Now for C there are various options for a verb. One option defines a verb in terms of certain primitive acts. This would be the conceptual dependency theory which goes back to the American psychologist Roger Schenck. An alternative representation would use thematic roles in order to define what sort of noun phrases can surface as a subject, as the object of a particular verb. So this means it's a verb of mental transfer which involves then certain roles and objects and classes to be associated with it. Well that is our determiner. It is demonstrative. Well we could say it's definite as well. Well in our preposition well we've seen that already is locative and directional so these are the semantic features associated with our preposition. Let's summarize and take our entry bill. Well here you see the collection of lexical specifications that are associated with our lexeme bill. Lexemes by the way I forgot to mention are always presented in capital letters. This is a convention in linguistics. Now here we have the phonology. This would be the morphology, syntactic representation and finally we have the sort of well semantics or meaning representation associated with our lexical entry. In the theory of grammar linguists are primarily concerned with those aspects that are necessary to define the legitimacy of syntactic structures. That is they confine themselves by and large to the morphological and syntactic aspects where a little bit of the semantics enters into the syntactic representation because we have to specify the arguments of verbs and of other categories in terms of thematic information. But that's a different story which will be explained in the e-lecture lexical insertion.