One of my readers posted a comment on this blog recently asking about the demonstrative/pointing lexical class of Inuktitut. This set is one of the most complex and extensively detailed lexical class in the Inuit language. I wrote a paper on it when I was starting out as a linguist and have had much time to think about it since. I must admit that I'm still trying to get my head around it for there is much more to it than first blush suggests.
For instance, this class is the only one with a prefix [ta-] in the whole of Inuit language, but more: it also has its own set of case endings (to, from, through, etc.).
Roughly following Dorais' and Mallon's analysis we may create a basic declension like this:
una/uvva 'this' manna 'here'
pinna/pikka 'that (up) there' panna '(up) there'
inna/ikka 'that one' anna/avva 'there'
kanna 'that (down) there' unna '(down) there'
but I treat another form and separate it from the subclass above:
majja majja manna piaktuq sikuup qulaa
'this here surface of the ice is slippery'
pagga pagga tuktu pangalikpuq qimaajuq
'that caribou (up) there is fleeing'
agga agga inuk maunga sanguniarmujuq
'that person (over) there will turn this way'
kagga kagga tiitaqarmat aiksilaurit
'that house (down) there has some tea so please go and get some'
and label it as "pointing to something with a continuous quality' or 'a specific aspect of something' - such as the slipperiness of the ice - or as "present, progressive verb emphasis" - as in, 'that airplane which is landing now'. In other words, this class differs from the first in that the first subclass refers to things as whole units while the second subclass refers to aspects or qualities of something.
Now, the interesting thing is that we may add the prefix [ta-] (Dorais calls this "thou" which I really like) to any one of the above:
and make the pointing/locating reference in relation not to the speaker but the one being spoken to.
But there is an additional complication because the allomorphic variations are not that simple and obvious when the prefix [ta-] is added on:
ta+una becomes taanna
ta+uvva becomes tavva
And I can't even begin to explain
taipsuma '(done) by that one there' -though I suspect the form comes from ta+uuma = ta + 'with respect to this one here'. But where does the -ips- come from? There are others like this, especially when we add case endings to any one of the forms:
taikuuna 'by way of that' could come from ta+ikka+na =thou+there+through
As I said, [-kkut] 'through' is manifested as [-na] as this whole class has its own case endings:
'from' [-ngat] instead of [-mit] as we'd normally say;
'through' [-na] instead of [-kkut] as we'd normally say;
'to' [-nga] instead of [-mut] as we'd normally say; etc.
but tauvani 'in that area' where the case ending [-ni] though it is obviously from [-mi] 'in', may only occur in the plural form of the case for this class, as in: Iqaluit becomes Iqalungni because the root "Iqaluit" is in the plural number.
I think figuring out the phono-morphemic rules here would provide some insights as to how irregularity in conjugations and 'strong' vs 'weak' verb forms arise. But I must admit that I'm currently stumped. We don't even know how the sole prefix in Inuktitut came into being, but I suspect it comes from 'third person possessive' evolving into 'second person possessive' for this particular class. I think figuring all this out would help us construct a theoretical framework that not only captures what is happening in Inuktitut but for all languages (though the motivations would differ the structural tendencies would be brought out more clearly.