Essentially, in the English language how many adjectives, nouns, adverbs, verbs, etc. could be combined to make the language possess a single word for a term. E.g. huge, giant, enormous, and big could all just go under big.
The thing about languages is that they can all say the same things. If it can't say something, it hasn't come across the concept yet. And when it has, it will make up or borrow a word to fill the gap. As a result, there will always be at least one "open" class of words, "open" meaning it can accept new members, like the classes "noun", "verb" and "adjective" in English.
This is not to say that those particular classes must be open for all languages, just that at least one of them must be (or even a different one entirely, like prepositions, one of English's closed classes - although I don't know of any examples of that...).
In a language with a closed verb class, for instance, new "verbs" are one of the native verbs with a noun or adjective complement, basically "I do a run" or "I am a runner" instead of "I run". If nouns are closed, new "nouns" are a native noun with a verb or adjective complement, "one who paints" or "painting one" for "painter", for example.
That last example also highlights another way that word classes can be "closed": inflection and derivation. Consider the case of English adverbs. Under a very niche definition that I just made up, this is a closed class: The only way new adverbs show up is when a new adjective shows up, and the productive -ly suffix turns it into an adverb. New adverbial concepts either take an adjective and stick on -ly or take a noun and stick it into a prepositional phrase.
Side note: this is what Esperanto - the person - was trying to do with Esperanto - the language: turn word class into an inflectional paradigm. The problem is, he was inconsistent. A broso is a brush - a thing that brushes - but a kombo is an act of combing, not a thing that combs. This is not to say that this is impossible, but you do have to be careful.
With all that out of the way, there's another problem. Using your example: while big, huge, giant and enormous all denote (that is, mean) the same thing, they connote (that is, imply) different things. "Enormous" is bigger than "huge" is bigger than "big". "Giant" is around "huge", but also implies that what is big isn't normally big, and sees use as a noun (meaning giant human). If you want to get around this, you'll need inflectional paradigms to change connotations, and connotation is a much wibblier concept than denotation - and denotation is plenty wibbly, which is why we even have different words for different things in the first place!
But if you're going to ignore all of that, then what you have is a pidgin at best, and an Orwellian nightmare at worst.
This question is probably more suited for the English stackexchange, where it would be promptly closed (or at least that's my gut feeling).
However, there is a way to estimate the number: the (English) WordNet groups words into synsets, each synset having a precise, pinpointed meaning (and covers a lot of vocabulary). So, for each of the parts of speech, just divide the number of words (literals) by the number of synsets to get the ratio of "words per meaning". For example, adjectives have the ratio 1.62.
However, there is a catch: words often mean more than just one thing, English is especially (for a European language) rich in homonymy. So, e.g. the noun giant has 7 different meanings, some of them perhaps differing only slightly, but some of them meaning very different things. Thus if you are after a quantitative number, you have to account for this somehow (would you count giant once or 7 times?).
And of course, you are opening a Pandora's box, because the deeper (beyond what WordNet covers) you go into less frequent words and more specialized terminology, the more and more unexpected meanings will surface and the numbers won't tell you anything anymore.