Defining words in text-to-speech

Let us begin by saying uncontroversially that we can assume that words which are different in orthography, phonetics and meaning are in fact different words. So if we have  

Again it is intuitive to think of these as two different words they mean different things, are spelt differently and are used differently. Pairs such as these are termed homophones and as with the homograph example, it is only a historical quirk that these have the same spoken form, they are not related in any other way. Consider now the following  

As with bear and bare these have different written forms but the same spoken form. But should we really regard them as different words In general this is inadvisable as in doing so we wouldn t be able to make use of the fact that apart from written form, these act identically in all other ways. The essential difference between the doctor cases and the bear and bare is in the meaning. in one case the meaning is the same while in the other it is different. Rather than count these as different words we say the different text forms are written variants of a single word, doctor (we can list more such as Dr., dr. and so on). A word such as DRIVE, which shares some written variants with doctor is considered a different word because it differs in meaning and spoken form  

Thus words can be considered homographs if only one of their written forms is the same. A single word can also have phonetic variants, and this is found in words such as either and economic  

In general, written variants are more common that pronunciation variants because they are usually the result of creativity on the part of writing practices rather than aspects of the language itself 

---