Re: Most developed conlang

On Wed, Apr 18, 2007 at 09:39:53AM -0700, David J. Peterson wrote:
[...]
> Dirk wrote:
> <<
> I don't know about Tokana, but I can tell you that Tepa has a lexicon
> of 600+ stems. Of course, that doesn't mean a thing, since the
> derivational processes in place will yield many more words than that.
> >>
>
> I echo this sentiment.  Consider a conlang like English that has verbs
> and a /po-/ prefix that's identical to the English /-er/ suffix.  It's
> not original, but it can be done in language, and can be done in a
> conlang.  Now say that language has 300 verbs.  Automatically, it has
> at least 600 words, whether they're written down or not.  And what
> about verb conjugations?  Using Esperanto as an example, or "estas",
> "estis", "estos", "esti" and "estus" all different  words?  If so,
> what about a language like Spanish, with even more conjugations?  If
> not, what are those things?  And what indeed does one do about
> Inuktitut or Turkish?
[...]

I agree with this as well. Counting words is fraught with many perils
once you cross the boundary of a single language. I've run into this
before when I wanted to evaluate how "up to speed" Tatari Faran was in
terms of completeness of lexicon. When you're comparing word counts
within a language, it makes sense, because the standard is common. But
once you compare different (unrelated) languages, all kinds of problems
come up:

- What exactly constitutes a "word"? Assuming that you want to
  extrapolate from word count to expressibility, you'd ideally want a
  "word" to correspond with "one unit of meaning". Unfortunately, the
  definition of a "word" is far from clear-cut. For example, Mandarin
  speakers consider that one word == one syllable, even though there are
  multisyllables that are properly only one semantic unit (e.g.,
  hou2die2). Yet these are often listed as separate words in
  dictionaries. What highly-inflected languages consider as a "word" may
  carry a lot more information than a "word" as defined by Mandarin
  speakers - they could convey tense, aspect, mode, position, state,
  etc., that count as separate "words" in Mandarin. Should different
  inflectional forms count as separate words, or not?

- Do affixes, suffixes, infixes, and other morphemes, count as
  individual "words"? What if they are not predictable (e.g., in
  inflecting languages where different morphemes are used with different
  words to express the same operation)? Where does one draw the line
  between affixes that count and affixes that don't? Every language has
  productive/non-productive processes (e.g., -ment in English), but, as
  Payne so eloquently says, nothing is completely productive, but there
  are different degrees of productivity. E.g., the -s plural suffix in
  English is very productive, but not completely (e.g., goose -> geese,
  sheep -> sheep). Some suffixes aren't very productive, but are common
  enough to be predictable in meaning (e.g. -ment).  Others are not
  predictable (re- in "release"). Where does one draw the line? Where
  would one draw a line that would be consistent *across languages*?

- What category of words "count" as contributing to the expressiveness
  of the language? E.g., grammatical particles such as "a", "the", "so",
  don't contribute very much to how much the language can express. If a
  language has plenty of these particles, should they count toward its
  expressiveness? What if the language has a higher word count by virtue
  of these semantically lacking (or empty) words than another language
  which has more "content" words, and which can therefore express more?
  This is a significant issue in Tatari Faran, where there are roughly
  as many "complements" as there are verbs, but these complements do not
  contribute significantly to how much the language can express since
  they are almost always synonymous with the verb, and are present only
  as a predicate reaffirmation device.

- And, as has already been mentioned, oligosynthetic languages may have
  a relatively low word count, but they are capable of expressing a lot
  more than an isolating language with around the same word count.

- Supposing that it is actually possible to come up with a standard that
  meaningfully compares word count across languages. Then there is the
  question of how units of meaning in different languages compare with
  each other. For example, conlang A could have a verb meaning "to sit",
  and conlang B has verbs meaning "to sit on a rock", "to sit on a
  chair", and "to sit on the floor", but lacks a general verb "to sit".
  Then one may argue that conlang A is more expressive, since it can
  express such concepts as "I sat on the hilltop", whereas conlang B
  would be unable to express such a thing without adding a new verb,
  even though in terms of word count, conlang B beats conlang A.

- Then there is the question of how expressive the language's
  grammatical structure is. For example, conlang A may have 5000 words
  (by some agreed on definition of "word"), and conlang B may have only
  3000 words. However, conlang A has no way of expressing comparisons
  ("X is more than Y"), whereas conlang B does. Which is more
  expressive? If you happen to need to express a lot of comparisons,
  then conlang B would seem more expressive; but if you only wanted to
  express lots of different concepts, then conlang A has a richer
  vocabulary to draw from. Of course, this is a contrived example, since
  a conlang lacking in comparatives would probably be considered
  incomplete in its grammar, but there are often innovative constructs
  in conlangs that are hard to compare with other conlangs in a fair
  way.


--T