Herman Miller wrote:

> I've been wondering about the possibility of taking some of my existing
> languages and developing their historical background to make them more
> realistic and less artificial-seeming.

The quick answer to the subject question is: yes, but with great difficulty, 
and much imagination. Consider: from modern Engl. (or any single modern IE 
language), would we be able to reconstruct anything resembling Proto IE? 
Engl. + Spanish? Swedish + Hindi? it would be an interesting 
thought-experiment... Proto-Austronesian was originally reconstructed based 
on just 3 modern languages (but with a lot of peeking at--er, background 
knownledge of-- several others), and even then Dempwolff didn't get it quite 
right (in the 1930s there was almost no reliable info on the Formosan 

> The problem with taking an existing language to start with is that I
> have to come up with a historical explanation for each feature of the
> language, or modify it in such a way that I can more easily explain it.

That's my problem with Kash; it was already carved in stone by the time I 
considered its history. Somewhere there's a big chart with the hypothetical 
preceding stages going back presumably 5-7000 years, but to fill it out 
would require, in effect, inventing a vast number of other languages and 
dialects (living and dead)-- though I could just come up with the 
phonologies of those stages. Grammar would be another matter. Then there's 
the problem of vocabulary loss/replacement and semantic change.

OTOH after devising a few Gwr words a priori, I decided I needed a 
proto-lang., so I made one (phonology, no grammar yet), and started making 
rules. That was about last Spring; I stopped work in Aug-Oct due to my 
sister's visit, and have only just now started to look at it again. AND I 
HAVE NO IDEA WHAT I WAS DOING!!!...aargh...though I think/hope the momentum 
can be recaptured.

Interestingly, given the nature of the modern languages, Kash has to have 
developed from a much more phonologically complex proto-stage; while 
proto-Gwr was much simpler than modern Gwr (e.g. 9 modern vowels + many 
diphthongs and triphthongs versus 3 proto-vowels /a i u/ and 4 diphthongs 
/-aw,-ay,-iw,-uy/) (And my proto-Gwr is only equivalent to, maybe, Italic in 
the IE schema; so there are several other as yet unknown branches of an even 
earlier proto-lang.)

> For instance, when I examined the tone patterns of two-syllable words in
> Simīk, I noticed that a few patterns were much more comman than the
> others, which could be explained by development from a simpler tone
> system in earlier versions of the language.

Varying stress patterns in the original? Prosody? (I'm only 1/2 familiar 
with how tones developed in Asian languages, which Gwr imitates; know 
nothing about the African or other tone systems)

But not all tone patterns
> fit into that system, so I had to assume they were borrowed from some
> other language, or were different in some other way (having a special
> tone pattern that was used for emphasis).

It could be that surrounding consonants could cause tonal variation; maybe 
the morphology? also dialect borrowing, the catch-all culprit.
> In the long run, is it better to start with one or more artificial
> proto-languages and develop them forward through time (which involves a
> lot of work on features that may not even make it into the future
> language system), or to start with an existing language and develop a
> history for it?

Starting with a proto-lang. is probably more desirable and realistic; and as 
you create the rules, you'll get ideas as to how they could differ in other 
branches; but that's probably too late for your purposes since you already 
have several completed languages. Assuming some of them are related, it 
should be possible to figure out some immediately preceding stages, then 
extrapolate to even earlier stages.
> For a specific example, I thought of taking Tirelat and trying to
> develop a history for it. Tirelat is a very regular and artificial
> language, which may actually be a result of engineering a more natural
> language to eliminate irregularities.

Is there a full description of the phonology, and a reasonably large vocab. 
we could look at?  A few years back 4 of us (IIRC) did a sort of 
reconstruction-game; I got Elliot Lash's Silindion and some very 
well-thought-out relatives; David Peterson got Kash and a bunch of very 
ad-hoc created relatives; but it worked, even though there were loose ends 
and mysteries remaining at the end.

But to start with something
> simple, the vowel system: Tirelat has 7 vowels /a e i @ 1 o u/, which
> may be long or short. Diphthongs do exist, but only /ai/ is common;
> combinations like /ia/ and /ui/ can be analyzed as consonant + vowel
> (/ja/, /wi/), except for the fact that there is no /ji/ or /wu/. So
> where do these 7 vowels come from?

Umlauting? Stressed vs. unstressed? Pre-tonic vs. post-tonic position? Older 
diphthongs? Lost consonants (things like h,?,G et al.) resulting in vowel 
coalescence? etc. etc.
> One thought that might explain the long vowels is that earlier versions
> of the language had more diphthongs, which simplified to single vowels.
> Notably, /a i u/ are more common than the other vowels. But any patterns
> in the data would have to be coincidental at this point, since the
> Tirelat vowel system wasn't developed with a history in mind.

Neither was Kash, so it takes a lot of acrobatics sometimes to figure out 
what a proto-form was. OTOH, I had to discard several of my original Gwr 
words because they were impossible derivatives (or else "dialect borrowing", 
but I really hate to resort to that).

 So in a
> case like this, would it be better to come up with an arbitrary history
> that doesn't fit all the facts (e.g. proto-language /o/ develops into
> /@/ except in the vicinity of a bilabial consonant, where it remains
> /o/, after which other phonetic changes occurred which caused /f/ to
> merge with /x/ resulting in a phonemic distinction between /x@/ and
> /xo/)? As it happens, /xo/ is slightly more common than might be
> expected, but /p@/ and /b@/ do exist, which can't be explained by the
> hypothetical sound changes.

Keep at it!! You'll come up with something.