Lexicon and Rules

To finish off this subsection, we need to briefly consider the work of Noam Chomsky, as it has, in a certain sense, provided the inspiration for the basic idea behind this foray into philosophy - that there is a close connection between the system concept and its pervasiveness and a certain feature of the mind [3]. Chomsky developed a rigorous description of a language in terms of its grammar, consisting of a lexicon and a set of rules, with the latter subdivided into a syntactic... 

Ostwald (1897) was aware of the fact that his rule was tenuous, since it was not based on a very large set of observations. In addition, if the metastable region were to shrink to a vanishingly small value, then sufficient time would not be allowed for crystallization of Form I to appear, and the rule would be invalidated. In fact, this does happen in many cases. Nevertheless, Ostwald s Rule has remained in the lexicon of crystal chemists, probably because it is generally observed that if a succession of polymorphic forms is obtained, those which appear later are generally more stable than those whieh appear earlier. 

One simple rule based systems works as follows. Firstly, we require a lexicon which gives the orthography for each word. With this, we can easily determine cheek whether two words in the lexicon have the same orthography, and from this compile a list of tokens which have two or more word forms. Next we write by hand a set of rules which fire (i.e. the activate) on the presence of the trigger token, for example ... 

Given this large offline lexicon, we see then that the real debate about rules vs lexicons is not one of quality, but rather one of balance between run-time and off-line resources. If we take the case where we include the entire off-line lexicon in the system lexicon, we will have a system which uses a considerable amount of memory, but where the processing speed is minimal (simply the small amount of time taken to look up a word). If on the other hand we create a system lexicon that is only a small subset of the offline lexicon, this will result in a smaller footprint, but as the pronunciation of absent words will have to generate at run-time, the processing costs... 

It is common in linguistics to say tiiat words are not atomic entities with regard to syntax or semantics but rather are composed of morphemes. If we consider the words big, bigger and BIGGEST we see there is a clear connection between the three words. The issue that concerns us is whether we should consider all of these as separate words, and store them in the lexicon, or just store the root form BIG and then use rules to derive the other forms. 

Pagel, V., Lenzo, K., and Black, A. Letter to Sound Rules for Accented Lexicon Compression. In Proceedings ICSLP 1998 (1998). 

---