Frislander's IE-lang scratchpad

A forum for all topics related to constructed languages
User avatar
Frislander
mayan
mayan
Posts: 2088
Joined: 14 May 2016 18:47
Location: The North

Re: Frislander's IE-lang scratchpad

Post by Frislander »

I've just realised I've missed out the ruki rule (I seem to have a habit of forgetting that. This would give the following i-stem paradigm (I've also realised I ave no idea where the zero-marked instrumental comes from so that's been restored):

Code: Select all

    SING    PLUR
NOM mɑ̃tiʂ   mɑ̃təj
ACC mɑ̃tim   mɑ̃tin
INS mɑ̃tiː   mɑ̃tip
DAT mɑ̃təjiː mɑ̃tip
GEN mɑ̃tiːʂ  mɑ̃tiːm
Suddenly I have more of a reason to keep this declension class separate, however I am still tempted to do some remodelling. We can import the instrumental singular -t from the o-stems, or alternatively the -p from the -eH2 stems, and then replace the dative singular with the locative as before, and leave it there fore now.

Code: Select all

    SING     PLUR
NOM mɑ̃tiʂ    mɑ̃təj
ACC mɑ̃tim    mɑ̃tin
INS mɑ̃tiːt/p mɑ̃tip
DAT mɑ̃tiː    mɑ̃tip
GEN mɑ̃tiːʂ   mɑ̃tiːm
Similarly the u-stems. We'll use huːnuʂ "son" as our example:

Code: Select all

NOM huːnuʂ
ACC huːnum
INS huːnuː
DAT huːnəwiː
GEN huːnuːʂ
LOC huːnuː
This is pretty much the same as the i-stems above, and once again I think I will have the locative replace the dative and the instrumental gain the -t from the o-stems or -p from the -eH2 stems. Note that the neuter θən "knee" is zero-marked in the nominative/accusative singular.

Now for the plural.

Code: Select all

NOM huːnəw
ACC huːnun
INS huːnup
DAT huːnup
GEN huːnuːm
LOC huːnuʂ
No alteration needed here aside from the loss of the locative. Note that the neuter shows the syncretic θənuː in the nominative/accusative here.

So for u-stems the paradigm is

Code: Select all

    SING      PLUR
NOM huːnuʂ    huːnəw
ACC huːnum    huːnun
INS huːnuːt/p huːnup
DAT huːnuː    huːnup
GEN huːnuːʂ   huːnuːm
Finally a note on -uH stems like huːʂ "pig".

Code: Select all

NOM huːʂ
ACC huːɑ̃
INS huːp
DAT huːiː
GEN huː
LOC huː
The locative is definitely being lost, but other than that this is probably just going to remain an irregular declension for now.

Code: Select all

NOM huː
ACC huːɑ̃
INS huːp
DAT huːp
GEN huːɑ̃ː
Again, not much to say here for now, just lots of irregularity and syncretism.

Code: Select all

    SING  PLUR
NOM huːʂ  huː
ACC huːɑ̃  huːɑ̃
INS huːp  huːp
DAT huːiː huːp
GEN huː   huːɑ̃ː
Birdlang
greek
greek
Posts: 631
Joined: 25 Dec 2014 20:17
Location: Virginia

Re: Frislander's IE-lang scratchpad

Post by Birdlang »

I’ve been wanting to make an IE-Lang all my life. Can you give me any tips? Anyways, cool, reminds me of Sanskrit, and also Pashto a little. Have you made a romanization yet?
Also, is there a book you’d recommend to help me. I’m trying to make both a kentum and satem language. And descendents of both.
User avatar
Frislander
mayan
mayan
Posts: 2088
Joined: 14 May 2016 18:47
Location: The North

Re: Frislander's IE-lang scratchpad

Post by Frislander »

I think today we shall first tackle r/n-stems. We'll look at wət "water" first, as it is definitely going to survive and can probably afford to be a little irregular.

Code: Select all

NOM wət
INS wətɑ̃ː
DAT wətəniː
GEN wətɑ̃ːh
(Note the presence of a full/o-grade in all forms by analogy). This seems like the locative is definitely being lost as usual, but I also fancy extending the instrumental again, though this time there's less reason for me to choose the -t rather than the -p suffix, so I'll leave it up in the air until I've finished with all these different declensions.

Now for a noun with both singular and plural forms jəkət (note I'm assuming the final -t of Sanskrit is original in some way for the moment).

Code: Select all

NOM jəkət
INS jəknɛː~jəkɑ̃ː
DAT jəkniː
GEN jəknəh~jəkɑ̃t
LOC jəkni
On again there is no good reason to retain the locative. The varying forms of the instrumental and genitive are dependent on whether there is an e-grade or zero-grade in the suffix respectively. My current inclination is towards the e-grade forms, but first let us consider the plurals:

Code: Select all

NOM jəknɛː~jəkɑ̃ː
INS jəknəp
DAT jəknəp
GEN jəknɛːm
LOC jəknəh
I wasn't entirely sure how to form the plural of this subclass, but I assume all-nasal is correct. Again the locative will be lost without trace, and there will be dative-instrumental syncretism. My current suspicion is that the e-grades will win out in the singular but the zero-grade in the nominative/accusative plural to maximise distinctiveness. The instrumental singular may then get a further extension on top of this. I also think I will keep the nominative singular -t as, while it is redundant, it is not ambiguous with any other form and also serves as an additional helpful marker of this noun's irregularity.

Just quickly for comparison we will consider the other version of this stem-class, represented by uːtkɛː "udder".

Code: Select all

    SING   PLUR
NOM uːtkɛː uːtkɑ̃ː
INS uːknɛː uːknəp
DAT uːkniː uːknəp
GEN uːknəh uːknɛːm
This is effectively identical except for the nominative singular, which looks very similar to the -eH2-stem. However otherwise the similarities are few, and not grounds for collapsing the declensions per se. However, the rarity and abject irregularity of this declension may give grounds for moving at least a few of the nouns involved into other declensions, and in this case I am tempted to do so, creating in the process what is in effect a class of neuter -eH2-stems. There could still be room for dialectal variation based on whether it is the uːtk- or uːkn- stem which is generalised: I will assume for now that the standard language shows the -tk- form.

Code: Select all

    SING    PLUR
NOM uːtkɛː  uːtkɛːh
INS uːtkɛːp uːtkɛːp
DAT uːtkiː  uːtkɛːp
GEN uːtkɛː  uːtkɛːm
So this leaves us with the following paradigm for "liver", and probably also for θəkət "shit" as well:

Code: Select all

    SING   PLUR
NOM jəkət  jəkɑ̃ː
INS jəknɛː jəknəp
DAT jəkniː jəknəp
GEN jəknəh jəknɛːm
Next time: finally tackling these -r and -n stems.

Birdlang wrote: 27 Oct 2018 23:28 I’ve been wanting to make an IE-Lang all my life. Can you give me any tips? Anyways, cool, reminds me of Sanskrit, and also Pashto a little. Have you made a romanization yet?
Also, is there a book you’d recommend to help me. I’m trying to make both a kentum and satem language. And descendents of both.
Wow, that's quite a bit of work there! I think you'd probably be best off working off of a handbook. I'm using Sihler's Comparative Grammar of Greek and Latin coupled with some of James Clackson's Indo-European Linguistics: an Introduction, as well as Wikipedia a bit, but really you should be consulting a wide variety of sources as there's no one set text. Furthermore I would strongly advise getting familiar with a few of the daughter languages, especially the ones you most want your respective kentum and satem languages to most resemble, because that will inform your choice of word-shapes quite a bit more than even PIE itself, because quite a bit of what's reconstructed for PIE (especially nominal ablaut) is more an explanatory device for the varying reflexes found in the daughter languages that can't be explained by regular sound laws (see Zekoslav and me's discussion of the -i stems for example).
User avatar
Frislander
mayan
mayan
Posts: 2088
Joined: 14 May 2016 18:47
Location: The North

Re: Frislander's IE-lang scratchpad

Post by Frislander »

OK, let's see if I can finish these consonant stems now (probably not, I'll have to save a draft most likely and come back to it). Let's start with the n-stem θuː "dog".

Code: Select all

NOM θuː
ACC θunɑ̃
INS θunɛː
DAT θuniː
GEN θun
LOC θ(w)ɑ̃ː
Standard procedure: lost the locative, contaminate the instrumental. For now the genitive looks like it'll experience some contamination, but let's look at the plural first.

Code: Select all

NOM θun
ACC θunɑ̃
INS θ(w)ɑ̃p
DAT θ(w)ɑ̃p
GEN θunɛːm
LOC θ(w)ɑ̃h
It looks like we'll have to reshape the genitive singular, and the instrumental singular, but to what? Let's take a look at uʂɛːn "pack animal" < "ox":

Code: Select all

    SING  PLUR
NOM uʂɛːn uʂən
ACC uʂənɑ̃ uʂənɑ̃
INS uʂɑ̃ɛː uʂɑ̃p
DAT uʂɑ̃iː uʂɑ̃p
GEN uʂɑ̃   uʂɑ̃ɛːm
From this the pattern is clear: the instrumental/dative plural should have the form -ɑ̃p not -əp. Also for "dog" we will add the o-stem genitive to avoid homophony with the nominative plural, and add to the instrumental as usual. This gives the following paradigm for n-stems:

Code: Select all

    SING     PLUR
NOM θuː      θun
ACC θunɑ̃     θunɑ̃
INS θunɛːt/p θunɑ̃p
DAT θuniː    θunɑ̃p
GEN θunəh    θunɛːm
Now for the r-stems. We'll take pətɛː "father" as our example word.

Code: Select all

NOM pətɛː
ACC pətəɾɑ̃
INS pəʂɛː
DAT pəʂiː
GEN pəʂ
LOC pətər
As we can see the -r- of the stem is almost lost, but does have funny effects on the preceding consonant in the oblique cases. This alternation may end up being mostly levelled out (with again corresponding restoration of the o-stem genitive), however for kinship terms I think it likely that it will stick around a bit longer. Ignore the locative as before.

Code: Select all

NOM pətəɾ
ACC pətəɾɑ̃
INS pətəp
DAT pətəp
GEN pəʂɛːm
OK, this gives us more support for levelling, since the syllabified rhotics don't affect the preceding consonant in the same way. Still, however, I think we shall retain the alternation for at least the kinship terms, giving the following paradigm.

Code: Select all

    SING     PLUR
NOM pətɛː    pətəɾ
ACC pətəɾɑ̃   pətəɾɑ̃
INS pəʂɛːt/p pətəp
DAT pəʂiː    pətəp
GEN pəʂ      pəʂɛːm
OK, and now let's just tackle the root nouns. We'll look at nək "night" as our example word.

Code: Select all

NOM nək
ACC nəktɑ̃
INS nəktɛː
DAT nəktiː
GEN nək
LOC nək
The genitive is probably being restored and there's no reason to keep the locative again. Probably I'll restore the genitive from o-stem contamination again.

Code: Select all

NOM nək
ACC nəktɑ̃
INS nəkpi
DAT nək(p)
GEN nəktɛːm
Yeah, there's definitely going to be some levelling, the instrumental/dative plural will take an oblique affix, extend the oblique and restore the o-stem genitive. however I'm not sure how to restore the nominative plural. Perhaps we shall look at pɛːt "foot" for inspiration:

Code: Select all

    SING  PLUR
NOM pɛːt  pət
ACC pətɑ̃  pətɑ̃
INS pətɛː pətəp
DAT pətiː pətəp
GEN pətəh pətɛːm
The answer is simple then: we'll extend the long vowel in the nominative singular to nək and other root nouns which did not show a lengthening to preserve the contrast, until the paradigms are reshaped again as will inevitably happen.

Code: Select all

    SING      PLUR
NOM nɛːk      nək
ACC nəktɑ̃     nəktɑ̃
INS nəktɛːt/p nəktəp
DAT nəktiː    nəktəp
GEN nəktəh    nəktɛːm
So I think that covers all the main stem types, next time I'll just go over a couple of irregular nouns before moving onto the new stuff.
User avatar
Frislander
mayan
mayan
Posts: 2088
Joined: 14 May 2016 18:47
Location: The North

Re: Frislander's IE-lang scratchpad

Post by Frislander »

Finally I'll just cover a couple of the irregular set of IE nouns before I move onto adjectives. I won't discuss these in detail since I've basically covered all the points I've wanted to cover so far. The first will be the mono-syllabic diphthongal stem nuːʂ "boat".

Code: Select all

    SING      PLUR
NOM nuːʂ      nɛːw
ACC nɛːwɑ̃     nɛːwɑ̃
INS nɛːwɛːt/p nuːp
DAT nɛːwiː    nuːp
GEN nɛːwəh    nɛːwɛːm
Next we have ɾiːʂ "riches" (note this is a mass noun).

Code: Select all

NOM ɾiːʂ
ACC ɾɛːjɑ̃
INS ɾɛːjɛːt/p
DAT ɾɛːjiː
GEN ɾɛːj
Next we have Tjuːʂ the name of the sky god (I realised that the rule as I'd written it implies that the post-vocalic glide merges with the vowel first, which is why this is the form not tiːwʂ as I imagined it to be previously. The feminine form however would still be Tiːwɛː).

Code: Select all

NOM Tjuːʂ
ACC Tiːm
INS Tiwɛːt/p
DAT Tiwiː
GEN Tiwəh
I think I'll pull a Latin here and use the oblique stems to form the word for "sky" and remake the "sky-god" paradigm accordingly, giving a regular neuter o/s-stem tiw for "sky" and a u-stem Tjuːʂ for the sky-god along the lines of "pig", like so.

Code: Select all

NOM tiw
INS tiwɛːt
DAT tiwiː
GEN tiwəh

NOM Tjuːʂ
ACC Tjuːɑ̃
INS Tjuːp
DAT Tjuːiː
GEN Tjuː
Finally there's hɛːwiː/hɛːwɛːɾ/hɛːwɛː "sun", where the -l- in the stem has screwed everything up completely (note I'm assuming a stage where the noun was assimilated to an ablauting paradigm like that found in r-stems).

Code: Select all

NOM hɛːwiː/hɛːwɛːɾ/hɛːwɛː
ACC hɛːwəjɑ̃/hɛːwəɾɑ̃/hɛːwɑ̃
INS hɛːwjɛːp/hɛːwɾɛːp/hɛːwɛːp
DAT hɛːwjiː/hɛːwɾiː/hɛːwiː
GEN hɛːwjəh/hɛːwɾəh/hɛːwəh
With this multiplicity of forms, I think it is likely that the ɕəʂ dialect forms with -ɾ will win out, simply because they preserve the case contrasts the best.

Oh and I also forgot n-stem neuters. We'll give as an example nəmɑ̃ "custom" (ablaut has been levelled out).

Code: Select all

    SING    PLUR
NOM nəmɑ̃    nəmɛːn
INS nəmnɛːt nəmɑ̃p
DAT nəmniː  nəmɑ̃p
GEN nəmnəh  nəmnɛːm
We will next cover adjectives, before we consider creating new cases.
User avatar
Frislander
mayan
mayan
Posts: 2088
Joined: 14 May 2016 18:47
Location: The North

Re: Frislander's IE-lang scratchpad

Post by Frislander »

I have an idea now to reshape the r-stems in the same manner as Vedic Sanskrit to something like -rs not -ros, which would give the following paradigm:

Code: Select all

    SING     PLUR
NOM pətɛː    pətəɾ
ACC pətəɾɑ̃   pətəɾɑ̃
INS pəʂɛːt/p pətəp
DAT pəʂiː    pətəp
GEN pətəʂ    pəʂɛːm
This will give further strength to the levelling of the paradigm with -t- in all forms.
User avatar
Frislander
mayan
mayan
Posts: 2088
Joined: 14 May 2016 18:47
Location: The North

Re: Frislander's IE-lang scratchpad

Post by Frislander »

OK, it's time for some adjectives damnit! I'm gonna assume that the late-PIE three-gender system is still operative. As in IE, adjectives follow nominal inflection, with the simplest type being o-stem for masculine/neuter and eH2-stem for feminine, like so, taking nəw "new" as our example:

Singular

Code: Select all

    MASC   FEM    NEUT
NOM nəw    nəwɛː  nəwɑ̃ː
ACC nəwɑ̃ː  nəwɛːm nəwɑ̃ː
INS nəwɛːt nəwɛːp nəwɛːt
DAT nəwiː  nəwiː  nəwiː
GEN nəwəh  nəwɛːh nəwəh
Plural

Code: Select all

    MASC   FEM    NEUT
NOM nəwɛːh nəwɛːh nəwɛː(h)
ACC nəwɑ̃ːh nəwɛːm nəwɛː(h)
INS nəwiːh nəwɛːp nəwiːh
DAT nəwəp  nəwɛːp nəwəp
GEN nəwɛːm nəwɛːm nəwɛːm
It is likely this, out of all the factors, that will likely prompt the expansion of the -t instrumental. We will also consider the u-stems, taking huːtuʂ "happy" (originally "sweet"). This might be a context where I'm tempted to keep the iH2-stems as a feminine marker (though see below):

Singular

Code: Select all

    MASC    FEM       NEUT
NOM huːtuʂ  huːtwiː   huːt
ACC huːtum  huːtwiːm  huːt
INS huːtuːt huːtujɛːt huːtuːt
DAT huːtuː  huːtujiː  huːtuː
GEN huːtuːʂ huːtujɛːh huːtuːʂ
Plural

Code: Select all

    MASC    FEM       NEUT
NOM huːtəw  huːtwiːh  huːtəw
ACC huːtun  huːtwiːm  huːtəw
INS huːtup  huːtujɛːp huːtup
DAT huːtup  huːtujɛːp huːtup
GEN huːtuːm huːtujɛːm huːtuːm
The nature of the phonological changes leads to the loss of -ro as a separate adjectival derivation: all such adjectives, such as ɾuʂ "angry, fierce" (from the same root as ɾuːtk- "red").

I have an idea, however, for a potential restructuring of the adjectival system. I haven't covered this yet, but as in Uralic the language will grammaticalise possessive suffixes, which may free up the genitive for alternative uses, and one I'm currently entertaining is freeing it up to make it a kind of particle linking modifier and head, like so:

Old: nəwɛːt jukɛːt "with (the) new yoke", New: nəw-əh jukɛːt of the same meaning.

Furthermore, I can also see the seeds of a collapse of the gender system, since abstract nouns will take the eH2-stems, so this may end up becoming a marker of abstract nouns rather than feminine gender. As part of this therefore this may further permit the case system simplification which I already have planned, and would spell the end of iH2-stems for good.

For that, though, we will need to look at pronouns.
User avatar
Frislander
mayan
mayan
Posts: 2088
Joined: 14 May 2016 18:47
Location: The North

Re: Frislander's IE-lang scratchpad

Post by Frislander »

So before I go onto pronouns, I've been thinking ahead a little to the verbs. I'm currently considering an Arapaho/Tocharian-type shift of the unmarked present indicative to a "non-affirmative" context, used with negatives and other things, while the derive present in *-ye/o- comes to mark the new present, and later by extension other indicatives, due to its collapse with a following vowel. Note that original *-ye/o- verbs with velar initials show the palatal glide fusing with the preceding stop, and will be reshaped accordingly. Similarly the "optative" (from thematic *-oy-) will end up as effectively a potential. We'll look at the verb pəɾ- "to carry": the non-affrmative includes the negative prefix in brackets.

Code: Select all

   IND     NON-AFF    POT
1s pəɾiː   (nə)pəɾɛː  pəɾəjɛː
2s pəɾih   (nə)pəɾəh  pəɾiːh
3s pəɾit   (nə)pəɾət  pəɾiːt
1p pəɾim   (nə)pəɾəm  pəɾiːm
2p pəɾik   (nə)pəɾək  pəɾiːk
3p pəɾjɑ̃ːt (nə)pəɾɑ̃ːt pəɾəjɑ̃t
User avatar
Frislander
mayan
mayan
Posts: 2088
Joined: 14 May 2016 18:47
Location: The North

Re: Frislander's IE-lang scratchpad

Post by Frislander »

OK, some pronouns. Pronouns are clearly much reduced in terms of case-marking vis-à-vis nouns: they only really distinguish nominative, dative and instrumental. Genitive enclitic pronouns have already become possessive affixes, and there is also a set of direct object clitics (derived from the non-tonic datives as well: the pronominal accusative in 1st and 2nd person has bee entirely lost), though not necessarily fused with the verb just yet. The paradigms for 1st and 2nd person are as follows:

Code: Select all

1st SING PLUR
NOM θɛː  wiː
DAT məp  ɑ̃hmiː
INS mət  ɑ̃hmət

2nd SING PLUR
NOM tuː  juːh
DAT təp  uhmiː
INS tut  uhmət
The instrumental forms are probably on the way out, for obvious semantic reasons.

Third person pronouns have also been grammatialised from demonstratives, while new demonstratives have been formed (this for another time). These undergo the same reduction of cases as the 1st and 2nd person ones.

Code: Select all

3rd m/ns    fs      m/npl   fpl
NOM hə/tət  hɛː     tiː/tɛː tɛːh
DAT təhmiː  təhiː   tiːp    tɛːp
INS təhmɛːt təhmɛːt tiːp    tɛːp
I'm assuming that the gender contrast survives, but it again seems ripe to be lost. It certainly will be lost in the following suffixes.

We will now turn to the clitics/suffixes. The direct object clitics are gammaticalised, as mentioned above, from the dative in the 1st and 2nd person, but for the 3rd person it's from the accusative, as evidenced by the nasal element:

Code: Select all

  SING PLUR
1 =miː =nəh
2 =tiː =wəh
3 =tɑ̃ː =tɑ̃ːh~tɛː
The variation in the 3rd person plural stems from whether the clitic is derived from the masculine or neuter. Note the lack of gender.

Similarly the possessive suffixes.

Code: Select all

  SING     PLUR
1 -məh     -nɛːh
2 -təh~təw -wɛːh
3 -təh~əh  -tiːʂɑ̃ː~iːʂɑ̃ː
Note that there is some variation, due to the potential homophony of the 2nd and 3rd person singular form. The -təw form derives from the equivalent free pronoun, while the t-less 3rd person forms are also derived from alternative forms without the prefix, also found in Latin.
User avatar
Frislander
mayan
mayan
Posts: 2088
Joined: 14 May 2016 18:47
Location: The North

Re: Frislander's IE-lang scratchpad

Post by Frislander »

OK, time to rejiggle the case system methinks.

As mentioned above, the current system of declensions seems ripe for some degree of simplification, though the earliest stages will likely still exhibit the system currently reached. Probably the most notable changes currently envisaged are the dative/accusative merger and the loss of the genitive. Given the results of the current set of sound changes, it seems likely that the merger will see the dative singular replace the accusative singular, while the accusative plural will remain and replace the dative plural, which in most declensions is syncretic with the instrumental plural anyway. This gives the following paradigms (again omitting the genitive singular; also note that I have chosen the p-variant of the instrumental outside of the nominative singular, for reasons I will explain below):

Code: Select all

    SING   PLUR
NOM wək    wəkɛːh
DAT wəkiː  wəkɑ̃ːh
INS wəkɛːt wəkiːh

NOM jukɑ̃ː  jukɛːh
DAT jukiː  jukəp
INS jukɛːt jukiːh

NOM wəkɛː  wəkɛːh
DAT wəkiː  wəkɛːm
INS wəkɛːp wəkɛːp

NOM nəp   nəpɛːh
DAT nəpiː nəpəp
INS nəpəh nəpəp

NOM mɑ̃tiʂ  mɑ̃təj
DAT mɑ̃tiː  mɑ̃tin
INS mɑ̃tiːp mɑ̃tip

NOM huːnuʂ  huːnəw
DAT huːnuː  huːnun
INS huːnuːp huːnup

NOM jəkət  jəkɑ̃ː
DAT jəkniː  jəknəp
INS jəknɛːp jəknəp

NOM θuː    θun
DAT θuniː  θunɑ̃
INS θunɛːp θunɑ̃p

NOM uʂɛːn  uʂən
DAT uʂɑ̃iː  uʂənɑ̃
INS uʂɑ̃ɛːp uʂɑ̃p

NOM nəmɑ̃    nəmɛːn
DAT nəmniː  nəmɑ̃p
INS nəmnɛːp nəmɑ̃p

NOM pətɛː  pətəɾ
DAT pəʂiː  pətəɾɑ̃
INS pəʂɛːp pətəp

NOM nɛːk   nək
DAT nəktiː nəktɑ̃
INS nəktɛːp nəktəp

NOM nuːʂ    nɛːw
DAT nɛːwiː  nɛːwɑ̃
INS nɛːwɛːp nuːp

NOM ɾiːʂ
DAT ɾɛːjiː
INS ɾɛːjɛːp

NOM hɛːwɛːɾ
DAT hɛːwɾiː
INS hɛːwrɛːp
New cases will in all likelihood be grammaticalised onto this new dative. I'm currently thinking that the instrumental/dative plural forms in -p will be significantly reshaped, and the wide distribution of the -p suffix in instrumentals makes it likely that this will be the form which wins out.

Now onto the potential list of new cases. Here there are some significant issues regarding distinctiveness, as some of the adpositions of PIE such as h₂epo "from" would merge through sound change with the already existing case endings (in this case the instrumental -əp). However, I have come up with the following set of postpositions-turned-cases: ət < *h2ed "allative", pəs < *bʰeǵʰ "without", əs < *h1eǵʰs "ablative", ɑ̃ːtɛː < *h1enter "inessive", kɑ̃k < *km̩th2 "comitative", ʂəh < *trh2os "perlative" , ɑ̃tkəɾ < *n̩dʰeri "subessive", upɛː < *uper "superessive". For an example we'll give a full paradigm of wək once more.

Code: Select all

    SING       PLUR
NOM wək        wəkɛːh
DAT wəkiː      wəkɑ̃ːh
INS wəkɛːt     wəkiːh~wəkəp
ALL wəkiːt     wəkɑ̃ːhət
SAN wəkiːpəs   wəkɑ̃ːhpəs
ABL wəkiːs     wəkɑ̃ːhs
INE wəkiːɑ̃ːtɛː wəkɑ̃ːhɑ̃ːtɛː
COM wəkiːkɑ̃k   wəkɑ̃ːhkɑ̃k
PER wəkiːʂəh   wəkɑ̃ːhʂəh
SUB wəkiːɑ̃tkəɾ wəkɑ̃ːhɑ̃tkər
SUP wəkəjupɛː  wəkɑ̃ːhupɛː
User avatar
Frislander
mayan
mayan
Posts: 2088
Joined: 14 May 2016 18:47
Location: The North

Re: Frislander's IE-lang scratchpad

Post by Frislander »

Oh, and some stuff on comparatives, seeing how I'm currently writing an essay on the topic for my supervisor.

So no-one's really sure we can reconstruct comparative inflections for PIE, but that's fine caus I'm gonna ditch them anyway. At least, I'm gonna ditch them as comparatives. See, there's good reason to suspect that the PIE suffix *-yos- from which comparatives in the daughter languages are based was actually an intensive, and that's brilliant for me, because the sound changes of the different forms result in some rather disconcertingly different forms, notably "long" tɛːk/ʂɛːɕəh~tiːɕəh~tɛːɕəh (damn *l really does screw everything up in this language doesn't it) and "strong"/"slow" kətuʂ/ʂətih (note, the slash doesn't represent the different reflexes, the two forms genuinely merge in both the normal and the intensive). This lets me have some fun with lexicalisation: I'm currently thinking for instance that "strong"/"slow" will split, with "slow" taking the normal form kətuʂ and "strong" the intensive ʂətih.

Comparison in this language therefore will have to be marked by some other means, which is easy enough: adpositions! I'm currently leaning towards the ablative əs, simply due to its relatively simple shape, though other options will be available and potentially different daughter languages could end up using different forms. I'm currently thinking comparand-standard-adjective ordering like so. In predicate constructions the "genitive" in adjectives will be dropped (I'm also thinking that in u-stems like huːtuʂ there will be a reanalysis of the -uʂ ending as part of the stem like the -ih in ʂətih and gain a -əh suffix when used attributively).

ɕənməh məpəs huːtuʂ
woman-1sgPOSS 1sg.DAT-ABL happy
My wife is happier than I am

wiː uhmiːs ʂətih
1pl.NOM 2pl.DAT-ABL strong
We are stronger than you
Last edited by Frislander on 27 Nov 2018 11:08, edited 1 time in total.
User avatar
Leo
rupestrian
rupestrian
Posts: 21
Joined: 28 Nov 2012 19:07

Re: Frislander's IE-lang scratchpad

Post by Leo »

Yes! New cases!
I'm confused why "we" is dative in your second example for expressing comparison.
User avatar
Frislander
mayan
mayan
Posts: 2088
Joined: 14 May 2016 18:47
Location: The North

Re: Frislander's IE-lang scratchpad

Post by Frislander »

Leo wrote: 26 Nov 2018 03:55 Yes! New cases!
I'm confused why "we" is dative in your second example for expressing comparison.
An error on my part, is has been corrected.
User avatar
Frislander
mayan
mayan
Posts: 2088
Joined: 14 May 2016 18:47
Location: The North

Re: Frislander's IE-lang scratchpad

Post by Frislander »

OK, some things on verbs, but before I do, one final sound change: I will now permit nasal vowels to merge with glides in the same manner as the non-nasal *a vowels, producing nasalised mid vowels. These will be fairly marginal, however they are relevant for today's discussion as they appear in the 3rd person subject markers.

As mentioned before, there is going to be some reshaping of the verbal system to allow for the creation of some more mood distinctions (because I reckon it's areally appropriate), and this will be done by creating a new regular indicative present from *-ye/yo- stems, and relegating other present tense formations to non-affirmative contexts (note that this means athematic verbs will likely end up being reshaped into thematics, which may prompt some spreading/merging of the athematic and thematic conjugations). Note that this is the regular *-ye/yo- formation giving suffixes in -i-: irregular forms produced by palatalisation of /k/ or other changes will show reanalysis, e.g. (nə)kɑ̃jɛː "come" < *gʷm̩joh2, reshaped affirmative kɑ̃jiː). So for the imperfective We have the following formations, taking wəhiː "wear", ɾiːkiː "leave", θəwiː "to hear"1.

Code: Select all

   AFFIR  NONAFF
1s wəhiː  (nə)wəhɛː
2s wəhih  (nə)wəhəh
3s wəhit  (nə)wəhət
1p wəhim  (nə)wəhəm
2p wəhik  (nə)wəhək
3p wəhẽːt (nə)wəhɑ̃ːt

1s ɾiːkiː  (nə)ɾinəkmi
2s ɾiːkih  (nə)ɾinəkhi
3s ɾiːkit  (nə)ɾinəkti
1p ɾiːkim  (nə)ɾinkməh
2p ɾiːkik  (nə)ɾinktə
3p ɾiːkẽːt (nə)ɾinkɑ̃ːt

1s θəwiː  (nə)θənuːm
2s θəwih  (nə)θənuːh
3s θəwit  (nə)θənuːt
1p θəwim  (nə)θənum
2p θəwik  (nə)θənuk
3p θəwẽːt (nə)θənõːt
The second person plural forms are derived from *-tH2-, with the laryngeal causing aspiration and later velarisation of the preceding stop.

Also note that the athematic nasal infix forms seen in (nə)ɾinəkmi are highly provisional, and are due for some reshaping themselves (particularly due to that retention of the full endings without cluster redution: it's just asking to have the paradigm reshaped), and I've a good mind to do the same to θənuːm and similar forms to something like the following:

Code: Select all

1s θənuː
2s θənuh
3s θənut
1p θənum
2p θənuk
3p θənõːt
This is effectively identical to having a thematic ending *-nwe- instead of an athematic *-new-, and I think I'll probably make this general as a result.

The language also preserves the optative in *-oy-, and like the *-ye/yo- suffix it has been extended across the paradigm, like so, using the same three roots as our examples (note the athematic formation of the 1sg).

Code: Select all

1s wəhəjɑ̃   ɾiːkəjɑ̃   θəwəjɑ̃
2s wəhəjəh  ɾiːkəjəh  θəwəjəh
3s wəhəjət  ɾiːkəjət  θəwəjət
1p wəhəjəm  ɾiːkəjəm  θəwəjəm
2p wəhəjək  ɾiːkəjək  θəwəjək
3p wəhəjɑ̃ːt ɾiːkəjɑ̃ːt θəwəjɑ̃ːt
Finally, a note on copulae. The formation of simple present *H1esmi "be" in the language ends up being so complex and indeterminate (mainly due to the issues arising from vowel deletion and clusters arising from athematic endings more generally), that I'm going to leave the language zero-copula in affirmative contexts, however with the accompanying particle *ne things become much simpler, so in light of the already existing Uralic tendencies in this language I'm going to give it a negative copula (potentially later a true negative verb) formed from a collocation *ne H1esmi and so forth.

Code: Select all

1s nɛːm
2s nɛːh
3s nɛːt
1p nɛːm
2p nɛːk
3p nɛːhɑ̃ːt
I'm happy to leave this form as-is, even if it is mildly defective, and leave any reshapings to the daughter languages.

1note that this present has also been majorly reshaped on the basis of the old present: it appears to reflect a formation along the lines of *ḱléwyoH2, however the regular reflex of that in the dominant ɕəʂ dialect would be something like *ʂuːyɛː; what appears to have happened is that a ɕək dialect form has been borrowed widely to remove the irregular alternation, and a new present has been formed from that.
shimobaatar
korean
korean
Posts: 10373
Joined: 12 Jul 2013 23:09
Location: UTC-04:00

Re: Frislander's IE-lang scratchpad

Post by shimobaatar »

I'm sorry it's taken me so long to actually comment on this. It's been on my "list", so to speak, since the first post was made.
Frislander wrote: 19 Oct 2018 15:09 I'm currently reckoning with the language being spoken somewhere eastwards, maybe round the Urals area?
Any idea how far north/south, approximately?
Frislander wrote: 19 Oct 2018 15:09 and that these people probably tagged along with the Iranians and/or Tocharians for a while before they split off north, hence several phonological commonalities can be found. Shared with Iranian we find satemisation, ruki and the complete merger of non-high vowels with accompanying palatalisation from *e. Shared with Tocharian we have the collapse of the three obstruent rows, though the exacts means and path this takes differs markedly from Tocharian.
I think I'm the least familiar with the "eastern" branches of the family, so this is very interesting. (That's not to say I don't personally find "western" IE conlangs interesting as well, but they're interesting in a different way.)
Frislander wrote: 19 Oct 2018 15:09 I reckon this language is fairly old, something like 2 thousand years BP, and I'm not certain whether I'll have a modern daughter or not.
If it were up to me, I probably would, because it's always interesting to see languages develop even further and fit into a changing world, but, of course, it's entirely your decision whether or not that's the direction you'd like to take this project.
Frislander wrote: 19 Oct 2018 15:09
Spoiler:
/p t k/
/s ʂ ɕ h/
/m n/
/w ɾ j/

/i iː u uː/
/ɛː ə/
/ɑ̃ ɑ̃ː/
I'm definitely a fan of this, particularly the three sibilants (+ /θ/ from the next post), single "liquid", and everything about the non-high vowels.
Frislander wrote: 20 Oct 2018 16:42 *l is lost, however there is variation as to whether it merges with *j or *r or is lost completely (the ɕəɕ/ɕəʂ/ɕək isogloss).
This is great. There's something about being able to name the different varieties after the form of a single word in each one that feels really "good" to me, for lack of a better word.

Looking at the wordlist below, it's even better knowing that the word means "wheel".
Frislander wrote: 20 Oct 2018 16:42 Aspiration came to be realised as strong velar frication with plosives, while aspirated affricates (from satemisation and palatalisation) became corresponding fricatives (so *tsʰ from *ǵʰ becomes modern /s/, *tɕʰ from palatalised *gʰ/gʷʰ becomes /ɕ/). Note that this is likely to have taken place after the second palatalisation, since we do not see a palatalisation of the velar element (so *tx from *dʰ never becomes *tɕ).

*tx become *tk, and velar frication is lost (resulting in a merger of *dʰ with unpalatalised *k, *kʷ, *g, *gʷ, *gʰ and *gʷʰ). This *tk cluster is simplified to /k/ when not inter-vocalic.
[+1] I love this.
Frislander wrote: 20 Oct 2018 16:42 Stress has at some point become almost completely word-initial, however it is post-initial if the first vowel is schwa and second is not.
Just to clarify, if the first two vowels of a word are both schwa, stress would be initial?
Frislander wrote: 20 Oct 2018 16:42 Coda -h in an unstressed syllable is lost, shortly followed by all word-final non-nasal short vowels (byebye o-stem nominative singular). Similarly unstressed word-initial schwa is lost. /h/ is also lost after a consonant word-finally.
Interesting how coda /h/ isn't lost universally. Looking at some of the word forms throughout the rest of the thread, coda /h/ is definitely part of what gives this language its unique aesthetic.

If the loss of a word-final non-nasal short vowel resulted in a new instance of coda /h/ in an unstressed syllable, would that /h/ then be lost, or would that rule already no longer apply? I'm wondering because of, for example, o-stem genitive singular forms like "wəkəh" and "jukəh".
Frislander wrote: 22 Oct 2018 23:31 Ok, let's make a start on this morphology shall we?
Frislander wrote: 23 Oct 2018 15:12 It's also a methodological thing - I'm currently working through the inherited morphology, seeing how the case system is likely to end up assuming no new stuff is grammaticalised, and doing so for each declension, before grammaticalising new stuff.
Clio wrote: 26 Oct 2018 19:12 I was looking forward to your post on the s-stems, and as expected it was a good one. I really enjoy reading through your thought process as you work through syncretism in and analogy from the output forms, especially here where they have a few really notable features (instrumental -p, tons of syncretism in the singular but none in the plural, etc.). When you derive the daughter languages, I'll be excited to see what happens to them.
I'm afraid I don't have very many specific questions or comments regarding the morphology posts, but I just can't overstate how much I love the way you're presenting things. It's very interesting, inspirational, and, frankly, fun to get to see your thought process as the language gradually takes shape. I really like that you're basing grammatical changes, at least in part, on how things have "naturally" wound up after the sound changes, and coming up with creative ways to repurpose things and use analogy to resolve potential issues that arise. I think I have a tendency to go into the process of deriving a language with a lot of predetermined ideas for how the resulting descendant "should" turn out, but to me, at least, this language feels like it's being created much more through "discovery", so to speak, than by following a rigid, preset plan, and I'm absolutely loving that.
Frislander wrote: 03 Nov 2018 18:11 Finally there's hɛːwiː/hɛːwɛːɾ/hɛːwɛː "sun", where the -l- in the stem has screwed everything up completely (note I'm assuming a stage where the noun was assimilated to an ablauting paradigm like that found in r-stems).
Spoiler:

Code: Select all

NOM hɛːwiː/hɛːwɛːɾ/hɛːwɛː
ACC hɛːwəjɑ̃/hɛːwəɾɑ̃/hɛːwɑ̃
INS hɛːwjɛːp/hɛːwɾɛːp/hɛːwɛːp
DAT hɛːwjiː/hɛːwɾiː/hɛːwiː
GEN hɛːwjəh/hɛːwɾəh/hɛːwəh
With this multiplicity of forms, I think it is likely that the ɕəʂ dialect forms with -ɾ will win out, simply because they preserve the case contrasts the best.
Looking back at the list of sound changes, there seems to be a rule that has *wr undergo metathesis to *rw. Does that not apply here because the /ɾ/ was originally *l, or because the cluster might have formed after the *wr > *rw change had already taken place? Or was /wɾ/ restored through analogy with the /wVɾ/ sequences in the nominative and accusative?

Spoiler:
Frislander wrote: 14 Nov 2018 19:26 So before I go onto pronouns, I've been thinking ahead a little to the verbs. I'm currently considering an Arapaho/Tocharian-type shift of the unmarked present indicative to a "non-affirmative" context, used with negatives and other things, while the derive present in *-ye/o- comes to mark the new present, and later by extension other indicatives, due to its collapse with a following vowel. Note that original *-ye/o- verbs with velar initials show the palatal glide fusing with the preceding stop, and will be reshaped accordingly. Similarly the "optative" (from thematic *-oy-) will end up as effectively a potential. We'll look at the verb pəɾ- "to carry": the non-affrmative includes the negative prefix in brackets.
Frislander wrote: 15 Nov 2018 18:36 We will now turn to the clitics/suffixes. The direct object clitics are gammaticalised, as mentioned above, from the dative in the 1st and 2nd person, but for the 3rd person it's from the accusative, as evidenced by the nasal element:
Frislander wrote: 15 Nov 2018 18:36 Similarly the possessive suffixes.
Frislander wrote: 19 Nov 2018 17:48 Now onto the potential list of new cases. Here there are some significant issues regarding distinctiveness, as some of the adpositions of PIE such as h₂epo "from" would merge through sound change with the already existing case endings (in this case the instrumental -əp). However, I have come up with the following set of postpositions-turned-cases: ət < *h2ed "allative", pəs < *bʰeǵʰ "without", əs < *h1eǵʰs "ablative", ɑ̃ːtɛː < *h1enter "inessive", kɑ̃k < *km̩th2 "comitative", ʂəh < *trh2os "perlative" , ɑ̃tkəɾ < *n̩dʰeri "subessive", upɛː < *uper "superessive". For an example we'll give a full paradigm of wək once more.
Frislander wrote: 23 Nov 2018 12:09 See, there's good reason to suspect that the PIE suffix *-yos- from which comparatives in the daughter languages are based was actually an intensive, and that's brilliant for me, because the sound changes of the different forms result in some rather disconcertingly different forms, notably "long" tɛːk/ʂɛːɕəh~tiːɕəh~tɛːɕəh (damn *l really does screw everything up in this language doesn't it) and "strong"/"slow" kətuʂ/ʂətih (note, the slash doesn't represent the different reflexes, the two forms genuinely merge in both the normal and the intensive). This lets me have some fun with lexicalisation: I'm currently thinking for instance that "strong"/"slow" will split, with "slow" taking the normal form kətuʂ and "strong" the intensive ʂətih.
Frislander wrote: 06 Dec 2018 13:11 Finally, a note on copulae. The formation of simple present *H1esmi "be" in the language ends up being so complex and indeterminate (mainly due to the issues arising from vowel deletion and clusters arising from athematic endings more generally), that I'm going to leave the language zero-copula in affirmative contexts, however with the accompanying particle *ne things become much simpler, so in light of the already existing Uralic tendencies in this language I'm going to give it a negative copula (potentially later a true negative verb) formed from a collocation *ne H1esmi and so forth.
[+1] Lots of very interesting ideas! I hope to see more of this language in the future.
User avatar
Frislander
mayan
mayan
Posts: 2088
Joined: 14 May 2016 18:47
Location: The North

Re: Frislander's IE-lang scratchpad

Post by Frislander »

shimobaatar wrote: 03 Jan 2019 17:06 I'm sorry it's taken me so long to actually comment on this. It's been on my "list", so to speak, since the first post was made.
Well I'll be honest this has fallen to the background recently, because I realised I had no idea what I do for descendants, so I reckon I'll start it again with the intent to make it a modern-day language instead, and I'll respond based on the ideas I currently have as well as what I've posted here before.
Frislander wrote: 19 Oct 2018 15:09 I'm currently reckoning with the language being spoken somewhere eastwards, maybe round the Urals area?
Any idea how far north/south, approximately?
Around the head of the Ob river, i.e. just north of the Altai mountains.
Frislander wrote: 19 Oct 2018 15:09 and that these people probably tagged along with the Iranians and/or Tocharians for a while before they split off north, hence several phonological commonalities can be found. Shared with Iranian we find satemisation, ruki and the complete merger of non-high vowels with accompanying palatalisation from *e. Shared with Tocharian we have the collapse of the three obstruent rows, though the exacts means and path this takes differs markedly from Tocharian.
I think I'm the least familiar with the "eastern" branches of the family, so this is very interesting. (That's not to say I don't personally find "western" IE conlangs interesting as well, but they're interesting in a different way.)
Thanks! I honestly find eastern IE to be more interesting too for similar reasons, though I will also admit that I realised I leaned perhaps a bit to heavily into the Indo-Iranian angle, so I'll probably change it to be more like Tocharian and be a centum language in the reboot.
Frislander wrote: 19 Oct 2018 15:09 I reckon this language is fairly old, something like 2 thousand years BP, and I'm not certain whether I'll have a modern daughter or not.
If it were up to me, I probably would, because it's always interesting to see languages develop even further and fit into a changing world, but, of course, it's entirely your decision whether or not that's the direction you'd like to take this project.
Well as I've said I'm probably going to make it a modern language next time, partly for the simple reason that I'm not sure how I'd work it so that the ancient language would even be attested, it's just not in the right place.
Frislander wrote: 19 Oct 2018 15:09
Spoiler:
/p t k/
/s ʂ ɕ h/
/m n/
/w ɾ j/

/i iː u uː/
/ɛː ə/
/ɑ̃ ɑ̃ː/
I'm definitely a fan of this, particularly the three sibilants (+ /θ/ from the next post), single "liquid", and everything about the non-high vowels.
Thanks! I think I'm actually quite pleased with the sibilants in particular, especially what ends up being the relative rarity of /s/ once /θ/ is added, which led to one of my few ideas for the daughter languages, that being the shift of /s/ to /ɬ/ à la Ugric, which could give some fun correspondences like ɬəmɑ̃ː ~ homō "person".

As for the non-high vowels I like what I did with those too, though tbh given the areal context I doubt the nasal quality of the low vowels would have remained for long given the absence of nasal vowels in Siberia, so for any reboot I probably won't keep them as being mainly nasal, though I might be tempted to say that they a nasalised before other nasals more easily perhaps.
Frislander wrote: 20 Oct 2018 16:42 *l is lost, however there is variation as to whether it merges with *j or *r or is lost completely (the ɕəɕ/ɕəʂ/ɕək isogloss).
This is great. There's something about being able to name the different varieties after the form of a single word in each one that feels really "good" to me, for lack of a better word.

Looking at the wordlist below, it's even better knowing that the word means "wheel".
Again thank you, I guess I just like isoglosses. And given how *kʷékʷlos is such a classic IE word I couldn't not use it for this. I knew I wanted to do something similar to Indo-Iranian and lose the *r-*l contrast and I had several different ideas so I though "sod it let's have all of them in there". That does pose further issues on the attestation side of things, like how such a language would even be attested with such variation, so if I'm honest this is probably something which has pushed me towards moving this language to the present day.
Frislander wrote: 20 Oct 2018 16:42 Aspiration came to be realised as strong velar frication with plosives, while aspirated affricates (from satemisation and palatalisation) became corresponding fricatives (so *tsʰ from *ǵʰ becomes modern /s/, *tɕʰ from palatalised *gʰ/gʷʰ becomes /ɕ/). Note that this is likely to have taken place after the second palatalisation, since we do not see a palatalisation of the velar element (so *tx from *dʰ never becomes *tɕ).

*tx become *tk, and velar frication is lost (resulting in a merger of *dʰ with unpalatalised *k, *kʷ, *g, *gʷ, *gʰ and *gʷʰ). This *tk cluster is simplified to /k/ when not inter-vocalic.
[+1] I love this.
I saw this change reading up on several North American languages ad I just kinda though "wouldn't it be cool if I could do and IE lang with this?", and I suppose it's also a natural way to lose an aspiration contrast in a part of the world where aspiration contrasts aren't common so I went for it. I'll probably modify it slightly if I'm going to redo this, probably to make *kx go to /x/, though that does raise some questions I've so far struggled to answer satisfactorily, like "why should *gʰ, *gʷʰ end up as fricative /x/ but *dʰ end up as a stop /k/?.
Frislander wrote: 20 Oct 2018 16:42 Stress has at some point become almost completely word-initial, however it is post-initial if the first vowel is schwa and second is not.
Just to clarify, if the first two vowels of a word are both schwa, stress would be initial?
Yes I think so.
Frislander wrote: 20 Oct 2018 16:42 Coda -h in an unstressed syllable is lost, shortly followed by all word-final non-nasal short vowels (byebye o-stem nominative singular). Similarly unstressed word-initial schwa is lost. /h/ is also lost after a consonant word-finally.
Interesting how coda /h/ isn't lost universally. Looking at some of the word forms throughout the rest of the thread, coda /h/ is definitely part of what gives this language its unique aesthetic.

If the loss of a word-final non-nasal short vowel resulted in a new instance of coda /h/ in an unstressed syllable, would that /h/ then be lost, or would that rule already no longer apply? I'm wondering because of, for example, o-stem genitive singular forms like "wəkəh" and "jukəh".
Exactly, the rule no longer applies. The o-stem genitives are from the full *-osyo suffix, so the *yo element would have been lost only after the coda-h had been lost.
Frislander wrote: 22 Oct 2018 23:31 Ok, let's make a start on this morphology shall we?
Frislander wrote: 23 Oct 2018 15:12 It's also a methodological thing - I'm currently working through the inherited morphology, seeing how the case system is likely to end up assuming no new stuff is grammaticalised, and doing so for each declension, before grammaticalising new stuff.
Clio wrote: 26 Oct 2018 19:12 I was looking forward to your post on the s-stems, and as expected it was a good one. I really enjoy reading through your thought process as you work through syncretism in and analogy from the output forms, especially here where they have a few really notable features (instrumental -p, tons of syncretism in the singular but none in the plural, etc.). When you derive the daughter languages, I'll be excited to see what happens to them.
I'm afraid I don't have very many specific questions or comments regarding the morphology posts, but I just can't overstate how much I love the way you're presenting things. It's very interesting, inspirational, and, frankly, fun to get to see your thought process as the language gradually takes shape. I really like that you're basing grammatical changes, at least in part, on how things have "naturally" wound up after the sound changes, and coming up with creative ways to repurpose things and use analogy to resolve potential issues that arise. I think I have a tendency to go into the process of deriving a language with a lot of predetermined ideas for how the resulting descendant "should" turn out, but to me, at least, this language feels like it's being created much more through "discovery", so to speak, than by following a rigid, preset plan, and I'm absolutely loving that.
Thank you. This tends to be how I've worked in my a posteriori projects mainly because that's kind of how these changes work in real life: sound shifts alter forms, speakers choose either to lose a form or find a way to repair the damage, and so on. Hence why I chose to use the scratchpad format, in order to represent this kind of thought process, since even the tiniest parts of a system can end up spreading out to most of a system, like how in Old English iirc the weak noun inflectiosn were most common until sound changes razed them all and then the few inflections from the comparatively small strong masculine declension were extended to most nouns. Furthermore I think it's also helpful when making an a posteriori conlang like this to walk through the changes which have gone into this kind of almost as they happen, because that's half the fun, and just presenting the final product and leaving the reader I guess to just work it out for themselves is less fun.
Frislander wrote: 03 Nov 2018 18:11 Finally there's hɛːwiː/hɛːwɛːɾ/hɛːwɛː "sun", where the -l- in the stem has screwed everything up completely (note I'm assuming a stage where the noun was assimilated to an ablauting paradigm like that found in r-stems).
Spoiler:

Code: Select all

NOM hɛːwiː/hɛːwɛːɾ/hɛːwɛː
ACC hɛːwəjɑ̃/hɛːwəɾɑ̃/hɛːwɑ̃
INS hɛːwjɛːp/hɛːwɾɛːp/hɛːwɛːp
DAT hɛːwjiː/hɛːwɾiː/hɛːwiː
GEN hɛːwjəh/hɛːwɾəh/hɛːwəh
With this multiplicity of forms, I think it is likely that the ɕəʂ dialect forms with -ɾ will win out, simply because they preserve the case contrasts the best.
Looking back at the list of sound changes, there seems to be a rule that has *wr undergo metathesis to *rw. Does that not apply here because the /ɾ/ was originally *l, or because the cluster might have formed after the *wr > *rw change had already taken place? Or was /wɾ/ restored through analogy with the /wVɾ/ sequences in the nominative and accusative?
I'll be honest I think one of my problems with the language in its current shape was that it's clear I wanted to have a fairly strict (C)V(C) language, but that I hadn't entirely worked out sound changes to resolve all of the clusters I still had left over, so I kinda had implied shifts like this one to get rid of word-initial *wr in words like *wreh2ds, but forgot to add them to my list of sound changes, and didn't apply them here. As a result I failed to specify that this changes was specifically word-initial, and hence would not have applied word-internally in the ɕəʂ dialects in the word for "sun".
[+1] Lots of very interesting ideas! I hope to see more of this language in the future.
Rest assured I will keep these ideas in any reboot, cause I enjoy these as well.
User avatar
Frislander
mayan
mayan
Posts: 2088
Joined: 14 May 2016 18:47
Location: The North

Re: Frislander's IE-lang scratchpad

Post by Frislander »

Time to discuss that reboot then. Let's start with sounds.

The consonant inventory is somewhat similar to how it was before, with a few notable differences.

/p t k/ <p t k>
/s ɬ ʂ ɕ x/ <s ł ṣ ś x>
/m n ɲ (ŋ)/ <m n ń ŋ>
/w l ɾ j ɣ/ <w l r j ɣ>

The reason the velar nasal is in brackets is because I haven't worked out sound changes that could produce it in native words, because original coda nasals were lost with nasalisation of the preceding vowel, so coda assimilation is impossible (to the point where clusters with first member nasal are pretty much only found in loans). All the others are found in native words through the application of regular sound changes, though some are made much more common through the presence of a large proportion of Ugric loanwords.

Pretty much the only allophonic variation to speak of is stops being somewhat voiced between vowels, though not necessarily consistently.

The history of this system is quite complex and multi-layered. As mentioned, the language is now centum and not satem, so there has only been one round of velar palatalisation, which produced /ɕ/ as the result. *s turned into a lateral fricative, so modern /s/ is essentially palatalised *t/*d plus loans (the same for the palatal nasal deriving from *n). The voiced velar fricative is the outcome of original *l, so the modern lateral is in native words the main outcome of non-initial *s, which remained as a voiceless lateral fricative in word-initial position and also as an outcome of both the clusters *st/*sd/*ts/*ds and also after original voiced aspirates, which as before ended up with stronger velar frication. The retroflex sibilant is basically from *Cr clusters. Basically the changes are a lot more involved this time around even compared to last time and really deserve their own post. The voiceless velar fricative, qua Ugric, is a reflex of an original dorsal before a back vowel at a certain pre-modern stage of the language; again the changes are complex and multi-layered and won't be fully explored here.

The vowels on the other hand are a complete c-change.

/ĭ i ŭ u/ <ĭ i ŭ u>
/e ə̆ ɵ̆ o/ <e ə ŏ o>
/æ ɑ/ <ä a>

Essentially, in native words, the low vowels and instances of ŏ that aren't directly adjacent to a velar stop /k/ are the old nasal vowels, or in the case of ŏ the syllabic nasals, and ä is especially uncommon in native words because it essentially is the nasalised form of an original long vowel. As before <ə e> are the reflexes of the merged PIE non-high vowels based on length, and as before the glides tend to coalesce with the vowels to produce diphthongs, and also vowels were rounded adjacent to old labiovelars, which produced instances of /k/ before rounded vowels, creating a contrast with /x/: c.f. the minimal pair kŏl- "to weep, wail" vs. xŏl "fir tree" - the latter is a loan, but as proof of principle here's the native xal "goose", which is indeed descended from *ǵhans.

So for some fun, here's some example derivations from PIE to this language, which I'm calling Xŏmsəńər Kele "marsh-island language".

*nókʷts > nŏx "night"
*péh2ur̩ > pew "fire"
*sál > łəɣ "salt"
*wĺ̩kʷos > wŏk "wolf"
*h3nómn̩ > nəm "name"
*kʷékʷlom > kŏkoɣ "wheel"
*h1ógʷʰis > ŏkś(əj) "snake"
*bʰérǵʰos > pśek "birch"
*ḱólHōm > kəɣä "grass"
*h2ŕ̩tḱos > əx "bear"
*déḱm̩t > səx "ten"
*km̩tóm> xŏta "hundred"
*h1éh1tr- > jeṣ "lung"
*dn̩ǵʰuh2 > tŏxu "tongue"
Salmoneus
MVP
MVP
Posts: 3033
Joined: 19 Sep 2011 19:37

Re: Frislander's IE-lang scratchpad

Post by Salmoneus »

Have you read up on the Ob-Yeneseian sprachbund for this? It seems like you'll be in that area...
User avatar
Frislander
mayan
mayan
Posts: 2088
Joined: 14 May 2016 18:47
Location: The North

Re: Frislander's IE-lang scratchpad

Post by Frislander »

Salmoneus wrote: 21 Jan 2019 13:37 Have you read up on the Ob-Yeneseian sprachbund for this? It seems like you'll be in that area...
Have you got any resources on this caus I can't seem to find anything on the Cambridge system, it's clearly not been much-worked on.
Salmoneus
MVP
MVP
Posts: 3033
Joined: 19 Sep 2011 19:37

Re: Frislander's IE-lang scratchpad

Post by Salmoneus »

No, I mention it literally just because somebody linked to this in another thread I was reading it a moment before seeing your post. It's a list of theorised sprachbunde that Uralic languages have been in.

Such an 'Ob-Yeneseian' or 'Ostyak' sprachbund would seem to be a possible influence on your language - but sadly he says that the scattered remarks about it in the literature still need to be systematised.

Also potentially of influence are the 'Core Uralic' area (which he sees as being influenced by neighbouring Indo-Iranian and Turkic languages, suggesting a broader zone that might be relevant to you), and the Ob-Ugric area.

He doesn't say much at all, but some of the citations may be interesting.
Post Reply