Pabappa/scratchpad

sound changes

Insanely complicated noun-verb couplings

08:50, 13 April 2022 (PDT)

In Play, noun-verb couplings like "the book you read to me" are an open class, with any noun able to bond with any verb. If Pabappa inherits the paradigm but loses the ability to create new words with it, the verb part could become an extremely large closed class, with perhaps hundreds of verbs able to pair with any noun, and the noun part would likely remain open. The sound changes would be the reason why the verb morpheme became a closed class, and these same sound changes would make the paradigm vastly more complicated and difficult to learn.

Ideally, this uses the archaic Play coupling system in which the noun stem is in the oblique case (with padded ʕ) and the infixed verb is also in the oblique case. This was already archaic in Play, but because Play scholars preserved their language with all its difficult aspects, it was still active and productive.

These would become a closed class early on in Pabappa, and this may have predated the split from Poswa, meaning that they would have become ossified even before the important sound changes took place.

An example of a word formed by this mechanism is the word for bottle.

Older ideas

May 9, 2021

Use -is as the Pabappa affix for "about, concerning, as for; like, similar to" (polysemy intended). This comes from Play -(t)efu and the loss of /t/ is because Pabappa uses the originally intervocalic form even after consonants stems since Pabappa has consonant-final reflexes for some originally vocalic stems. A possible exception is that /t/ could survive after /p/, since /pt/ > /t/.

Apr

Ape 2021

need to check semantic shifts on all words with "wall" in the original definition, as there are far too many and these must have come from an experiment with different forms

"square, tile" also occurs suspiciously often

Possession markers

Feb 8, 2021

Need to find out why -di is described as obsolete below.

Possible retention of thrown-back classifier suffixes

Feb 6, 2021 bumped 09:35, 4 February 2023 (PST)

NOTE: Since Play most likely does not have this, Pabappa cannot either unless an unusual type of analogy took place. It might be best to just let it go.

The very archaic classifier prefixes of Tapilula were deleted in Gold, but a few managed to survive by clinging onto the previous word in the sentence, where they behaved as suffixes. This word was usually, and perhaps always, the first word in the sentence because the prefixes were primarily used as verbal person and object class markers, and therefore appeared primarily in sentences with the structure SV.

Note that when the agent was 3rd person and the patient was a SAP, the object markers were further from the root than the subject markers were, and thus were unstable. Verbs with this inverted structure could be analyzed as passive verbs, but there were never any markers for voice on the verb; the person markers determined that.

Pabappa could hang on to these, such that e.g. inanimate nouns take a fusional suffix to mark that they are the agent of a sentence with a 1st or 2nd person patient, and perhaps Pabappa could in some situations even retain the 3rd person markers which are ultimately derived from the noun class prefixes of Tapilula.

This could lead to unusual situations such as the suffix -ta "man, agent, career marker" turning into -top whenever the action being performed requires the use of tools. Or perhaps /-top/ could simply come to be seen as a variant of /-ta/ used for careers that require tools. note that this is because tool is /ku/ and has nothing to do with the SAPs

If retained, these could be called dynamic classifiers.

Disyllabic verb endings

Apr 28, 2020

Still try to rework Pabappa's verb endings so that even though the tense marker is interior to the person marker (like Turkish), the outermost vowel still varies according to tense as in Poswa. Thus the resulting disyllabic verb ending has the same atomic feel that Poswa's does.

If Pabappa's 1st person singular accusative pronoun evolves from šuvi a bup >>>> op, it could appear as a Poswa-like suffix on 3rd person animate subjects, e.g. pisaniop nannabi "the thorn cut me", alongside a passive verb construction like /pisani nannampipu/.

Though the /-p/ might disappear before word-initial /m n/ and cause at least some vowel-initial words to get an /s/, depending on whether they had earlier begun with /š/ or /ž/ (the few rare primordial vowel-initials would likely have been assimilated by this time into the other two categories). /p#l/ probably survives rather than changing to /bl/. This loss of /-p/ before nasals is a learned behavior, because there was no /-op/ in this position at the time of the shift of /pn pm/ >> /n m/.

THE SECTION ABOVE THIS SHOWS A MUCH BETTER WAY TO MARK THE PERSON OF THE PATIENT ON THE AGENT, HOWEVER.

Body parts as agents

Apr 28, 2020

Unlike Poswa, possessed nouns in Pabappa all take third person verb agreement, even if they are body parts. To make instrumental forms, the instrumental case must be used. (I dont have a good idea yet for how to derive the instrumental case of the possessed form of a noun; it can only be done analogically, as the required proto-form didnt eixst.)

Dont derive a new instrumental affix from a verb.

Corrupt etymologies

Apr 28, 2020

A tiny but still real number of words in the Pabappa dictionary show loss of final /-di/ on the assumption that that was going to be the 2nd person nominal possession marker. Since it isnt, those words must be found and reverted. A few new analogies could be set up where /-pu/ drops instead, but these should probably be marked in the margins in case the ending needs to be changed again.

I had thought that the word for hip was one of these, but it seems to have never been added.

Pronouns and person markers

Apr 27, 2020

• Pabappa probably needs person markers on verbs because I cannot find a language that has nominal possession markers without also having verbal person markers. The verbal person markers will be suffixed after the tense marker and may or may not be cognate to the nominal possession markers. The third person verbal marker is probably Ø. The first and second person, if cognate to the nominal possession markers, may be either identical to them, or derived from the same stems but without the original final /-s/. But note that the final /-s/ may not have made a difference in the surface forms anyway.

• Deriving pronouns could be a first step to getting both verbal person markers and justifying the nominal possession markers.

• If the first person pronoun comes from Play šuvi a bu "(I) feel that (I) speak", i.e. "one speaking", it could evolve into just o or u in a standalone form but would be preceded by -ps- word-internally. There is no easy way to simplify this cluster.
• If instead it comes from Play šuvi a ba (effectively the same meaning as above), it would evolve into one of oa ~ ua ~ wa in bare form and to psoa ~ psua word-finally.

• The uncertainty above about the tonic vowel is because it is not certain that Pabappa analogized /u/ > /o/ in closed syllables the way Poswa did ... it is actually quite likely that it was not analogized, because Pabappa would have very little precedent for this shift whereas Poswa had a large number of verbs where /o/ appeared in both closed and open syllable forms of the verb.
• pu is another possibility. The path is

• Play šuvi a bu > pšulabu > pšužu > pšu > šu > hu > fu > wu (free) and pu (suffix). This would still happen only in the bare form, and so for it to be a suffix it would need to have been reordered at some stage, preferably towards the end when /f/ would have been seen as just an ordinary phoneme. The nominal possession marker could be an unetymological po or pa, or a more proper (but still reliant on analogy) -so ~ -sa.

• Similarly,

• Play šuvi a ba > pšulaba > pšuža > pšua > šua > hua > fua > wua (free) and pua (suffix), but there is no way to compress /wua/ > /wa/.

• Any form deriving from a bare /šuvi/ in Play is unlikely because it would be identical to a content word and would end in a consonant.
• Second person at first seems simpler, if the brightest path is taken:

• Play ma nuvi aa > mabulā > mabyža > mabʷa > maba in free form (note that it is not /mabu/), and
• Play ma nuvi aa > mabulā > mabyža > mʲyža > vyža > viza ~ voza > dia ~ doa, almost certainly with loss of /a/ due to some sort of analogy, likely then followed by stabilization as either /di/, /do/, or /da/. It is possible also that /vyža/ > /vža/ since it would only ever occur after a vowel, and then /vža/ > /vza/ > /va/ > /da/ directly. This calls into question whether /vz/ should be treated separately from /v/ and /z/, however, since the shifts of /v/ > /d/ and /z/ > /Ø/ are listed as simultaneous, when they were likely separated by at least a few decades.

• Yet another possibility is that /vz/ > /uz/ (since the /v/ was always labialized), and then /uz/ > /u/ > /w/. This would mean that the second person marker is always and only -wa.
• The 2nd person pronoun could be ta in order to avoid repetition like "wa .....-wa", even though this is etymologically improper. Alternatively, a new, suppletive 2nd person pronoun could arise for at least the nominative form.
• Could do /vʷz/ > /bz/ > /b/ instead, making the 2nd person marker -ba instead of /-wa/ or any of the others, which would in some verbs lead to a combined suffix /-baba/ which superficially resembles Poswa.