Loading ...
Sorry, an error occurred while loading the content.

Re: Portamenteau as a derivational process

Expand Messages
  • H. S. Teoh
    ... Heh... when I started replying I intended to just give a brief overview, but then I just couldn t resist giving all the juicy details afterwards. ... It s
    Message 1 of 18 , Apr 21 11:37 PM
    • 0 Attachment
      On Sat, Apr 20, 2013 at 11:40:30PM -0400, Alex Fink wrote:
      > On Fri, 19 Apr 2013 23:47:47 -0700, H. S. Teoh <hsteoh@...> wrote:
      >
      > >Alright. Believe it or not, I actually managed to sort through the
      > >entire TF lexicon! I haven't done a thorough analysis yet, but here
      > >are some interesting initial observations.
      >
      > Awesome. And if this is not thorough, I quail to imagine the thorough
      > version...

      Heh... when I started replying I intended to just give a brief overview,
      but then I just couldn't resist giving all the juicy details afterwards.
      :-P


      > >Verbs:
      > >- Conclusions:
      > > - While there are a good number of verbs derived with other affixes,
      > > they seem arbitrary and non-systematic.
      >
      > Hm, I wonder if this could show the "stem alternations" category:
      > proto-TF verbs, in whatever inflectivity they retained, had more than
      > one stem, with a lot of different lexical patterns of stem
      > correspondence.

      It's possible, I suppose. I wonder if this may have happened in many
      things other than verbs... for example, here are some cognates (or words
      that look suspiciously like cognates) of "blood":

      aisu (finalizer) bleeding, bloodily. Used with the verb _hutakas_,
      meaning to smite, strike with great force, give a beating.

      asai (finalizer) reddeningly, of a crimson, bloody color. Used with
      the adjective "red".

      saipia' (masc. n.) spear (_pia'_ means pole or shaft, so this suggests
      the derivation as "blood-shaft")

      saisu (fem. n.) blood, internal bodily fluids.

      akaisu (v) to bleed (_ak_ appears to derive from the postposition
      _aka_, "out", "out from")

      It's interesting that of the above words, some appear to derive from the
      stem _ais_, some from _sai_, with the noun for "blood" itself deriving
      from both. It certainly suggests some kind of ancient stem alternation.



      > >Nouns:
      > > - Vowels in feminine nouns show a weak tendency towards /i/, /ei/,
      > > and /u/, whereas masculine nouns lean towards /a/ (I didn't
      > > quantify this precisely, though).
      >
      > Hm, what is the historical basis of assignment to genders, for words
      > that didn't fall in the semantic core of any of the genders? This
      > could be a remnant of a phonological family resemblance that was used
      > to make those assignments.

      It's possible, for sure. The current gender assignments seem to be based
      on some kind of contrasting system (more on this later). Which may have
      been an extrapolation from an ancient more phonologically-based
      assignment system. It's interesting to note that the prevalent case
      particles have /a/ in the masculine and /ei/ in the feminine, whereas
      now masculine nouns appear to have a weak tendency towards /a/ and
      feminine nouns towards /ei/ (and /u/), thus suggesting that, at least in
      part, the ancient gender system may have been based on some kind of
      vowel harmony or rhyming scheme.


      [...]
      > >- Conclusions:
      > > - It looks like -ra vs. -ei, -i, -ri may be vestiges of an ancient
      > > gender inflection system. (There are also a few other traces not
      > > listed here, like _minas_ / _minein_ "name (masc/fem)".)
      >
      > Name of a male / of a female, you mean? Is there any cultural reason
      > why that's lexicalised?

      Heh, I haven't really thought about that. It's interesting to note that
      this distinction seem to be quite deeply-rooted in the language; for
      example, you have contrasting pairs like:

      hatse' (masc) a man's hair
      suiri (fem) a woman's hair

      dukun (masc) male servant
      hina (fem) female servant

      tsuna (masc) traditional men's garment
      sura (fem) robe (traditional women's garment)

      Many words for body parts are variable-gender, with the gender of the
      case particle chosen to match the owner's gender. (This may be a modern
      innovation, though; in some early TF notes I hypothesized that the case
      particles were ancient pronouns, so the variable gender of the body part
      words may have been merely a reinterpretation of what used to be a
      possessive pronoun.)

      Now, outside of biological gender, a good number of current masc/fem
      nouns appear to have some kind of contrasting assignment scheme:

      keras (masc) poisonous fruit
      ba'as (fem) edible fruit

      baran (masc) morning/day
      mubun (fem) night

      sinkan (masc) the daytime sky
      panikan (fem) the night sky

      bunai (masc) thumb
      jiri (fem) finger

      jukasu (masc) fast-moving, blocky lava (Hawaiian: a'a)
      sinasu (fem) slow-moving, "smoother" lava (Hawaiian: pahoehoe)

      kauna (masc) crow, predatory bird
      tsuinit (fem) songbird

      kuen (masc) a leafy tree
      tinka (fem) a coniferous tree

      maha (masc) liver (phys); heart (psych.)
      umpas (masc) heart (phys); (fem) temper, emotion (psych.)

      miri' (masc) ice, frost
      miris (fem) snow

      pasanan (masc) town
      misanan (fem) village

      batsiu insect; masc. for large insects, fem. for small insects


      So maybe the san faran just like categorizing things this way. I'll have
      to prod my informant for more explanations. ;-)


      > > - It seems that -n/-an is a likely nominalizing suffix, though the
      > > picture is muddied by a good amount of finalizers also showing the
      > > same suffix.
      > > * I should note also, that -n/-an is currently the productive
      > > genitive suffix; I'm not sure how this fact ties in with the
      > > above data.
      >
      > I can just about imagine a process whereby a nominaliser most
      > canonically used on quality stems ('long' -> 'a long one') jumps from
      > there to be used on possessors ('John' -> 'John's one'), and then
      > through use in apposition gets its start as a genitive suffix.

      I like this idea! It seems to jive with the fact that a lot of -n/-an
      occurrences appear to involve words of very basic meanings (and
      therefore probably of more ancient origin). The currently-productive
      -n/-an genitive suffix feels like it should be a more recent
      development, esp. since it doesn't exhibit many irregularities that
      ancient inflections are prone to exhibit.


      > >Finalizers:
      > >- Conclusions:
      > > - I'm not sure how to interpret the -n/-an derivations for
      > > finalizers. They seem a bit too frequent to write off as
      > > coincidence; yet they don't really fit in with -n/-an being
      > > primarily a nominalizing suffix.
      > > * Moreover, I had previously thought that they were obligatory
      > > adverbs that over time lost their semantic content, yet they
      > > don't seem to share the derivational morphology of adverbs (see
      > > below). Instead, they seem to related to nouns! So I really
      > > don't know what to make of this right now.
      >
      > Well, one could certainly try to tell some interesting stories about
      > why it would be nouns found in that position.
      >
      > They could have started out, for instance, as old, semantically
      > bleached objects. Maybe proto-proto-...-TF favoured syntactically
      > transitive verbs in some environments, and so semantic intransitives
      > took a generic nonreferential complement noun, the way that Mandarin
      > _chi1fan4_ 'eat (intr.)' = 'eat a meal, eat rice' has. (Maybe they
      > even fìrst got a start in negatives, in the French fashion of _ne
      > mange mie_, _ne bois goutte_, _ne marche pas_, etc.)
      >
      > Or, closer to your original idea, they could have started out as nouns
      > (maybe in some case form? though maybe a zero-marked one?) with
      > adverbial force. Thinking about why these might become so useful as
      > to become fixed, if not just through expressiveness, hm, perhaps they
      > had some sort of completive ~ resultative force ('walked _to the
      > [very] top_', 'burnt _to cinders_', 'killed _to a corpse_', etc.), and
      > passed through a period where they contributed perfectivity?

      I like this idea. Perhaps they were not a single class of words, but
      remnants of various words that occurred frequently with the same verbs
      until their semantic value has been bleached and they merged into an
      eclectic collection of verb-specific finalizers. The pairing of verbs
      with finalizers then took root, and eventually caused *all* verbs to
      pick up a finalizer via analogy. So you'd have ancient nouns, ancient
      adverbs, adjectives, and who knows what else classed as finalizers, all
      retaining remnants of diverse word forms as witnessed today.

      Also, the bit about not sharing the derivational morphology of adverbs
      may not represent the true situation accurately. Yes I didn't find
      significant evidence for shared *derivational morphology*, but then a
      good proportion of basic adverbs are irreducible, and so are a large
      core of finalizers. For all I know, many of the finalizers may have been
      reclassified adverbs. Those finalizers that show nominal-like morphology
      could have been derived from ancient nouns, like you said.

      Your idea also quite accurately describes the force of TF finalizers;
      you have examples like _hutakas aisu_: _hutakas_ = to strike with great
      force, to beat up; _aisu_ appears to derive from "blood", so you have
      what looks like the calcified remnant of an expression akin to "beat to
      a bloody pulp".

      Then a generic verb like _tapa_ (to walk, to go) that can be paired with
      multiple finalizers, among which are _anan_ (seems to be a cognate of
      _anui_ "upwards"), for "to go up". Perhaps _an_ was an ancient stem
      meaning "up" or "top", which gave rise to _anui_ (to go upwards, from
      the attested directional suffix _ui_), and _anan_ (with the -an
      nominalizer), so perhaps _tapa anan_ meant "to go to the top", and in
      the modern language _anan_ lost its use as a standalone noun, and got
      reclassified as a finalizer. A similar analysis would work for _tapa
      ta'an_ (to go down(wards)).

      This also poses an interesting question of how to analyze certain
      zero-valent expressions like: _peira ta'an_ (it is raining, from _peira_
      "(neut. n.) rain", and _ta'an_, the finalizer carrying the force of
      "down"). If _ta'an_ was a noun in proto-TF, then this expression is a
      bit puzzling; why put two nouns together? One possible explanation may
      be that it was a statement of location (TF has zero copula, so the
      ancient expression may have meant "rain is [arrived] here_on_the_ground"
      or something along those lines -- _ta'an_ may have meant "ground", or
      perhaps the nominalization of an ancient adverb "down", so "the place
      down below").


      > (AFMCL, Sabasasaj also has bipartite verbal stems. The historically
      > less verbal component is also difficult to explain and has a few
      > frustrating behaviours, but I do have written down that I imagine the
      > ancientest cases as being nominal complements or the first sort above,
      > that became fixed.)

      Would my idea of TF finalizers being a merger of words of diverse
      classes work in Sabasasaj as well? :) If it started out as expressions
      like "kill to_a_corpse", "eat to_satiation", as you said, then what
      stops the same process of semantic bleaching from applying to things
      like "run really_very_fast" or "talk very_wordily" or "reddened
      like_crimson", etc.?

      This is particularly compelling in TF's case, as predicative adjectives
      are *also* paired with finalizers, so you have expressions like:

      sura sei pirat inai.
      dress CVY:FEM yellow(ADJ) bright(FIN)
      The dress is yellow.

      sinkan sa karat asai.
      daytime_sky CVY:MASC red(ADJ) as_blood(FIN)
      The sky is red [like blood].

      The finalizers can easily have been adjective modifiers ("the dress was
      yellow -- bright yellow!"; "the sky was red -- like blood!") that got
      bleached of their semantic content. (In modern-day TF, the _inai_ in
      _pirat inai_ doesn't carry any factual meaning anymore; a dull yellow
      dress is still _pirat inai_.)


      > > - Reduplication seems to turn up with some frequency in finalizers.
      > > They don't seem to be elsewhere besides the currently productive
      > > intensification of adjectives (_baasa_ "big" -> _basa-baasa_ "very
      > > big"). I think it's safe to interpret them as the result of
      > > onamapoia, unrelated to the reduplication of adjectives.
      >
      > Reduplication is weird that way, historically, just showing up with no
      > particular antecedent -- maybe as you say from onomatopoeia, here, or
      > some other kind of affective process. There's no reason why proto-TF
      > mightn't've had two different productive reduplication processes.

      True. I sorta modelled many aspects of TF after the Austronesian
      languages, where reduplication appears to be a common trait; so proto-TF
      having multiple reduplication processes is well within my desired
      language parameters. :)


      > >Adverbs:
      > >- 23/43 = 53% of adverbs are irreducible.
      > [...]
      > >- Conclusions:
      > > - It looks like the directional suffix -ui is the most common
      > > adverbalizing suffix.
      >
      > It's worth asking what the semantics of these underived TF adverbs
      > are, and how internally cohesive they are as a class. Certainly, in
      > the Western grammatical tradition, 'adverb' has been a little bit
      > abused as a catch-all.

      Adverbs in TF are a very strictly-defined class (at least in modern-day
      TF): they are modifiers that immediately follow the main verb. They
      include things like "run *fast", "talk *loudly*", "jump *repeatedly*",
      etc..

      There are some decidedly non-IE adverbs, though, among which are what I
      call "adverbs of manner". These are verb modifiers that serve that
      function of what in English requires an auxiliary verb construction; for
      example:

      tapa ha buara na!
      walk(V) start(ADV) volcano RCP:MASC
      Start walking to the volcano!

      In the English translation, the main verb is "start" modified by
      "walking"; in TF, on the contrary, the main verb is "walk" and "start"
      is an adverb ("startingly"?). In the same vein:

      tapa bat buara na!
      walk(V) stop(ADV) volcano RCP:MASC
      Stop walking to the volcano!

      Quite a good number of constructions in English that requires an
      auxiliary verb are subsumed by adverbs-of-manner in TF:

      arap pera beira sei.
      pick_up(V) try(ADV) stone CVY:FEM
      Try to pick up the stone!

      tsana irei maharan so.
      talk(V) continue(ADV) story CVY:NEUT
      Continue to tell the story.


      > The TF class in -ui, I suppose, contain mostly adverbs describing
      > motion (paths? manners?)

      They are directional adverbs (in the above sense), for example:

      tapa baranui
      walk eastwards

      juerat ta'ui
      look downwards

      They form an interesting class, because they exhibit dialectal
      differences in TF. The san faran are divided into the northern tribes
      and the southern tribes, and there is a prominent volcano that roughly
      forms the boundary between north and south. So you have the adverb
      _buaranui_ (from _buara_ "volcano" + -n- [genitive? ancient
      nominalizer? something else?] + -ui), which to a Northerner means
      "southwards", but to a Southerner means "northwards". :)

      There's also _huaranui_, from _huara_ ("lake") + -n- + _ui_, which is
      only attested in Southern dialects, meaning "southwards" -- because
      there is a lake on the south end of Fara, but to a Northerner, the
      southern boundary is the _buara_; the lake further south is outside
      their territory and therefore of little interest.

      OTOH, the Northerners have a word _hiranui_, from _hirana_ (snow-capped
      mountain) + _ui_. The _hirana_ here refers to a specific, particularly
      spectacular stratovolcano with a snowcapped peak that lies beyond the
      barrier mountains along the edge of Northern Fara. This mountain is not
      visible from southern Fara, and so the Southerners have no such word. :)

      For east/west:

      _mubunui_: westwards (from _mubun_ "night" + _ui_)
      _baranui_: eastwards (from _baran_ "morning" + _ui_)

      Why west is associated with night seems a bit strange (the darkness also
      rises from the east, after all!), but perhaps in proto-TF "mubun"
      actually meant "sunset" and subsequently got shifted into "night".

      Then you have _ta'ui_ (downwards) and _anui_ (upwards), which have
      cognates that suggest the stems _ta'_ and _an_ for down and up,
      respectively.

      (There are probably other -ui words besides these, but I haven't
      coined^H^H^H^H^H learned about them from my TF informant yet. :-P)


      > >Finally, some interesting observations about the cross-category
      > >distribution of affixes (as before, this doesn't include cases which
      > >don't have unambiguous evidence that it isn't merely part of the
      > >stem):
      > >- -ai and -s occur in the most categories: verbs, nouns (all 3 genders),
      > > adjectives, adverbs, and finalizers.
      > >- -an occurs in nouns of all 3 genders, adjectives, and finalizers.
      > >- -n (including -an) is notably *absent* from all verbs.
      >
      > Of course, not all of these words need contain the derivational affix
      > it looks like they might. They could have these endings for other
      > (coincidental lexical) reasons. Or have you already accounted for
      > that, and all these words contain derivational affixes as attestable
      > by comparing other words containing their stem?

      I've already accounted for this; there are a great many other words that
      look like they have one of the above affixes, but which have no cognates
      that differ only in that affix, and so can't be proven to be formed by
      affixation (or equivalently, can't be proven not to have a stem that
      just coincidentally contains a suffix lookalike).


      > >- -su occurs in nouns of all 3 genders as well as one complement; in all
      > > cases it is clearly related to water (_usu_) or liquid (esp. blood,
      > > _saisu_, with cognates _aisu_, _akaisu_, etc.).
      >
      > Maybe an old liquids noun class, if the noun class system was once
      > larger; more likely, to my feeling, just an old root. Maybe _usu_ is
      > original, was reduced to _-su_ in some compounds after a V-final first
      > element through crasis, and then that form was generalised.

      Yeah, I'm leaning towards that explanation as well.


      > >So in conclusion, it looks like one may safely conclude that:
      > >- -ai/-itai are verbalizing suffixes which may still be productive in
      > > modern-day TF;
      > >- -ra, -ei/-i/-ri are likely vestiges of an ancient gender inflection
      > > system;
      > >- -n/-an are nominalizing suffixes.
      > >- -s/-as are adjectivising suffixes, and nouns having this suffix are
      > > possibly substantives;
      >
      > All of this looks solid.

      Yeah, I was frankly quite surprised by this. My initial impression was
      that the distribution of suffixes like -s or -n were too arbitrary, or,
      in some cases, "inverted" from what they should be. Which may still be
      the case for a (very) small number of words, but it turns out that the
      distribution is much more consistent than I had thought!

      Furthermore, my initial analysis seemed to suggest an uneven
      distribution of the nominalizing suffixes between noun genders, but this
      turned out not to be the case: the numbers were skewed by the fact that
      the TF lexicon didn't have equal numbers of masc/fem/other nouns. Once
      that was accounted for, the unevenness turned out to be statistically
      insignificant.


      > >Future directions of investigation:
      >
      > and looking forward to all of this!
      [...]

      Heh, on that note, I had just (re)discovered recently that I had
      actually composed a poem in TF many years ago, which I had never
      published. Due to some errors in the poem, I dug a bit deeper in my old
      files, and discovered to my mixed joy/chagrin that the poem was
      apparently constructed as a modernization of a sketched initial version
      written in "Old Tatari Faran"! Joy, because it was really neat to
      rediscover such an interestingly constructed poem; chagrin, because the
      "Old Tatari Faran" as represented by the sketch showed some features
      that are rather ... shall we say, difficult to reconcile with some of
      the evidence I gathered in the above analyses? And also poses some hard
      questions about how it could have plausibly evolved into modern TF.

      The good thing, of course, is that none of this was ever published (in
      fact, the notes containing the sketch explicitly stated that it was
      never to be published -- I had anticipated myself, apparently :-P), so I
      don't feel the pressure of having to revise a ton of TF material should
      I find that I need to drastically reshape what I think Old TF should
      look like.

      But I may even have anticipated this latter point as well, because the
      notes *also* stated that the sketch represented "an ancient poem in Old
      Tatari Faran, as recited by an old lady in such-an-such a village".
      (The unpublished poem was supposed to be a rendition of said ancient
      poem in modern-day TF.) Meaning that the language represented by the
      sketch may have been a fabrication or retcon of said old lady, and thus
      may not accurately represent what Old TF *actually* was. Gah. I may have
      outsmarted myself, 'cos looking over the sketch, I have a hard time
      remembering what I had in mind when I wrote it. :-/

      What do y'all do when you rediscover old conlang notes that had been
      forgotten for many years, and find that it contradicts what you had
      subsequently developed in your conlang?


      On Sun, Apr 21, 2013 at 07:39:53PM +0200, BPJ wrote:
      > 2013-04-20 08:47, H. S. Teoh skrev:
      > > - It seems that -n/-an is a likely nominalizing suffix, though the
      > > picture is muddied by a good amount of finalizers also showing the
      > > same suffix.
      > > * I should note also, that -n/-an is currently the productive
      > > genitive suffix; I'm not sure how this fact ties in with the
      > > above data.
      >
      > There is no law against homophonous suffixes, or suffixes
      > becoming homophonous, is there? ;-)
      [...]

      True! That did occur to me as well. After all, there seems to be signs
      of distinction between -s/-as for the adjectivising suffix, vs. -s/-is
      for the currently-productive partitive case suffix. It's very possible
      that there may have been two distinct -n suffixes in proto-proto-TF that
      became lookalikes due to mergers in some ancient vowels.


      T

      --
      We've all heard that a million monkeys banging on a million typewriters
      will eventually reproduce the entire works of Shakespeare. Now, thanks
      to the Internet, we know this is not true. -- Robert Wilensk
    • Roger Mills
      BTW, apologies for not going further with my field investigation.... Real Life and some bodily ailments got in the way. Bah. ... aisu    (finalizer)
      Message 2 of 18 , Apr 22 7:38 AM
      • 0 Attachment
        BTW, apologies for not going further with my "field" investigation.... Real Life and some bodily ailments got in the way. Bah.

        --- On Mon, 4/22/13, H. S. Teoh <hsteoh@...> wrote:

        aisu    (finalizer) bleeding, bloodily. Used with the verb _hutakas_,
            meaning to smite, strike with great force, give a beating.

        asai    (finalizer) reddeningly, of a crimson, bloody color. Used with
            the adjective "red".

        saipia'    (masc. n.) spear (_pia'_ means pole or shaft, so this suggests
            the derivation as "blood-shaft")

        saisu    (fem. n.) blood, internal bodily fluids.

        akaisu    (v) to bleed (_ak_ appears to derive from the postposition
            _aka_, "out", "out from")

        RM: looks immediately like -su might mean "(in a) liquid (state)-- as indeed you point out later in the post.  (Coincidentally, /su/ is the combining form of Kash /sau/  (written "sawu") 'water' and does indeed mean 'liquid').


        suiri    (fem) a woman's hair
        .......
        sura    (fem) robe (traditional women's garment)

        RM There's that su- again: might womanly things be considered more "liquidy ~  flowing than male things??

        Many words for body parts are variable-gender, with the gender of the
        case particle chosen to match the owner's gender. (This may be a modern
        innovation, though; in some early TF notes I hypothesized that the case
        particles were ancient pronouns, so the variable gender of the body part
        words may have been merely a reinterpretation of what used to be a
        possessive pronoun.)

        Now, outside of biological gender, a good number of current masc/fem
        nouns appear to have some kind of contrasting assignment scheme:

        keras    (masc) poisonous fruit
        ba'as    (fem) edible fruit

        baran    (masc) morning/day
        mubun    (fem) night

        RM nothing odd here-- le jour, la nuit, el dia, la noche etc. etc.

        sinkan    (masc) the daytime sky
        panikan    (fem) the night sky

        RM but why aren't these more visibly related to the foregoing?

        miri'    (masc) ice, frost
        miris    (fem) snow

        RM el hielo, la nieve etc

        pasanan    (masc) town
        misanan    (fem) village

        RM might pa- mean larger, or mi- mean smaller???

        batsiu    insect; masc. for large insects, fem. for small insects

        So maybe the san faran just like categorizing things this way. I'll have
        to prod my informant for more explanations. ;-)

        RM I'm not sure such things are amenable to analysis, even by a linguistically savvy informants. "That's just the way we say it..."

        True. I sorta modelled many aspects of TF after the Austronesian
        languages, where reduplication appears to be a common trait; so proto-TF
        having multiple reduplication processes is well within my desired
        language parameters. :)

        RM true dat. YOu can have full reduplication (pluralizing mainly, but also to give a sense of "something like (base)"; partial redup (1st syllable usually); Kash also allowed final syll. redup.


        There are some decidedly non-IE adverbs, though, among which are what I
        call "adverbs of manner". These are verb modifiers that serve that
        function of what in English requires an auxiliary verb construction; for
        example:

            tapa    ha         buara   na!
            walk(V) start(ADV) volcano RCP:MASC
            Start walking to the volcano!

        In the English translation, the main verb is "start" modified by
        "walking"; in TF, on the contrary, the main verb is "walk" and "start"
        is an adverb ("startingly"?). In the same vein:

            tapa    bat       buara   na!
            walk(V) stop(ADV) volcano RCP:MASC
            Stop walking to the volcano!

        Quite a good number of constructions in English that requires an
        auxiliary verb are subsumed by adverbs-of-manner in TF:

            arap       pera     beira sei.
            pick_up(V) try(ADV) stone CVY:FEM
            Try to pick up the stone!

            tsana   irei          maharan so.
            talk(V) continue(ADV) story   CVY:NEUT
            Continue to tell the story.

        RM but what does TF do when the "aux." is the main verb, as in an imperative:
         Start now! Don't walk there! Do go on (continue)!  etc.???


        They form an interesting class, because they exhibit dialectal
        differences in TF. The san faran are divided into the northern tribes
        and the southern tribes, and there is a prominent volcano that roughly
        forms the boundary between north and south. So you have the adverb
        _buaranui_ (from _buara_ "volcano" + -n- [genitive?  ancient
        nominalizer?  something else?] + -ui), which to a Northerner means
        "southwards", but to a Southerner means "northwards". :)

        RM this happens in a lot of Austronesian lgs. as well, where "sea(wards)" can mean east, west, north, south depending on the island/region; same with "inland"; and others

        What do y'all do when you rediscover old conlang notes that had been
        forgotten for many years, and find that it contradicts what you had
        subsequently developed in your conlang?

        RM I shake my head and go, whatever made me do that??? :-)))) Except in the case of Gwr, when I discovered several original forms that didnt comply with the rules I eventually formulated...........In that case, don't repaid, create anew....
      • Roger Mills
        Afterthought: Some of the finalizers look a bit like old serial verb constructions.
        Message 3 of 18 , Apr 22 7:41 AM
        • 0 Attachment
          Afterthought:

          Some of the "finalizers" look a bit like old serial verb constructions.
        • H. S. Teoh
          ... No problem, the offer is still open if you wish to take it up sometime. ... Yeah, -su is pretty consistent throughout. In fact, there s even _sumu _
          Message 4 of 18 , Apr 22 10:32 AM
          • 0 Attachment
            On Mon, Apr 22, 2013 at 07:38:26AM -0700, Roger Mills wrote:
            > BTW, apologies for not going further with my "field" investigation....
            > Real Life and some bodily ailments got in the way. Bah.

            No problem, the offer is still open if you wish to take it up sometime.
            :) My TF informant is pretty easy-going about these sort of things. ;-)


            > --- On Mon, 4/22/13, H. S. Teoh <hsteoh@...> wrote:
            >
            > aisu    (finalizer) bleeding, bloodily. Used with the verb _hutakas_,
            >     meaning to smite, strike with great force, give a beating.
            >
            > asai    (finalizer) reddeningly, of a crimson, bloody color. Used with
            >     the adjective "red".
            >
            > saipia'    (masc. n.) spear (_pia'_ means pole or shaft, so this suggests
            >     the derivation as "blood-shaft")
            >
            > saisu    (fem. n.) blood, internal bodily fluids.
            >
            > akaisu    (v) to bleed (_ak_ appears to derive from the postposition
            >     _aka_, "out", "out from")
            >
            > RM: looks immediately like -su might mean "(in a) liquid (state)-- as
            > indeed you point out later in the post.  (Coincidentally, /su/ is the
            > combining form of Kash /sau/  (written "sawu") 'water' and does indeed
            > mean 'liquid').

            Yeah, -su is pretty consistent throughout. In fact, there's even _sumu'_
            "well", probably _su_ + _mu'_ (hole), i.e., "water hole".


            > suiri    (fem) a woman's hair
            > .......
            > sura    (fem) robe (traditional women's garment)
            >
            > RM There's that su- again: might womanly things be considered more
            > "liquidy ~  flowing than male things??

            It's certainly possible! This seems to correspond with _jukasu_ vs.
            _sinasu_: blocky, chunky lava vs. smooth, ropy lava.


            [...]
            > keras    (masc) poisonous fruit
            > ba'as    (fem) edible fruit
            >
            > baran    (masc) morning/day
            > mubun    (fem) night
            >
            > RM nothing odd here-- le jour, la nuit, el dia, la noche etc. etc.

            True.


            > sinkan    (masc) the daytime sky
            > panikan    (fem) the night sky
            >
            > RM but why aren't these more visibly related to the foregoing?

            Well, I'm not sure where _sin-_ comes from, but _pani-_ seems to be
            cognate with _panis_ "star" and _panei_ "the Milky Way". So _panikan_
            seems to be something like "starry sky", or "the sky of stars".


            > miri'    (masc) ice, frost
            > miris    (fem) snow
            >
            > RM el hielo, la nieve etc

            True.


            > pasanan    (masc) town
            > misanan    (fem) village
            >
            > RM might pa- mean larger, or mi- mean smaller???

            Hmm. I couldn't find any obvious cognates, but a remote possibility
            might be _paan_ [pa:n] "wide", "expansive", "grand". Didn't find any
            obvious cognates of _mi_ either, but perhaps _mihat_ "to meet, to gather
            together"?

            Seems to be on rather shaky ground, but assuming these aren't just
            coincidences, the historical construction may have been:

            pasanan < paa(n) "grand" + san "person" + an (nominalizer)
            misanan < mi(hat) "to meet together" + san "person" + an (nominalizer)

            I.e. the assembling together of people = village; a wide, expansive
            [area/meeting] of people = town.


            > batsiu    insect; masc. for large insects, fem. for small insects
            >
            > So maybe the san faran just like categorizing things this way. I'll have
            > to prod my informant for more explanations. ;-)
            >
            > RM I'm not sure such things are amenable to analysis, even by a
            > linguistically savvy informants. "That's just the way we say it..."

            True. Plus, you may just get a folk etymology which may or may not have
            any basis in fact. :) (I remember making a fool of myself trying to
            explain the Hokkien term for mainland China /tŋ sũa/ as "long mountain",
            due to the homophonous /tŋ/ "long", but the *real* etymology is "Tang
            mountain" (i.e. of the Tang dynasty)).


            > True. I sorta modelled many aspects of TF after the Austronesian
            > languages, where reduplication appears to be a common trait; so
            > proto-TF having multiple reduplication processes is well within my
            > desired language parameters. :)
            >
            > RM true dat. YOu can have full reduplication (pluralizing mainly, but
            > also to give a sense of "something like (base)"; partial redup (1st
            > syllable usually); Kash also allowed final syll. redup.

            That's interesting. I haven't really heard of any langs with final syll.
            reduplication.

            In TF, the reduplicated stem gets reduced in some ways (long vowels get
            shortened, final consonants get dropped, etc.), with the original stem
            appearing last and always bearing the accent:

            baasa ['ba:sa] "large" -> basa-baasa [basa'ba:sa] "very large"
            (the vowel in the 2nd syll. is phonemically /a:/ reduced to /a/, but
            pronounced very laxly, like [V] or even [@] in practice).

            tsat ['ts)at] "fast" -> tsa-tsat [ts)a'ts)at] "very fast".
            (the final /t/ is dropped in the reduplicated stem)

            tihai ['tihaj] "old" -> tiha-tihai [tiha'tihaj] "very old"
            (/ai/ reduced to /a/ in the redup stem)

            I wonder if a hypothetical initial-stressed redup might have given rise
            to final syll. redup within TF phonotactics, maybe like:

            baasa > *baasa-basa > *baasa-sa

            This doesn't really work in TF, though, 'cos the modifier-final word
            order seems to prefer initial redup for prosodic reasons: an NP with an
            adjective has the form N - ADJ - CASE, and the case marker is (almost)
            always unstressed, and usually that means ADJ is stressed. Furthermore,
            between ADJ and CASE one can insert a demonstrative, which is also
            usually unstressed. Since TF is pitch-accented, that means the stressed
            syllable of ADJ is high-pitched, and everything that follows must be in
            descending pitch. That's a lot of syllables to cram into a descending
            pitch contour if ADJ starts having tail-redup. :)

            buta' baasa fi- sei
            [bu"ta? 'ba:sa fi sej]
            hut big DEM CVY:FEM
            That big hut

            (The secondary stress on _buta'_ is a bit crammed, and probably
            realized as mid-pitch as opposed to the usual high pitch.)

            buta' basa~baasa fi- sei
            [bu"ta? basV'ba:sa fi sej]
            hut very~big DEM CVY:FEM
            That very big hut

            (Having the redup stem in front allows room for secondary stress
            on _buta'_; it can be mid-high or even high pitch, followed by
            low-pitched _basa~_, followed by another high pitch peak on
            /ba:/, then trailing off to the end of the NP.)

            *buta' baasa~basa fi- sei
            *[buta? 'ba:sabasV fi sej]
            hut *big~very DEM CVY:FEM

            (5 syllables to squeeze into a falling pitch contour after /ba:/
            -- seems a bit of a stretch)

            Prosody is one of those things that I really like about TF (even though
            I do still need to iron out quite a few wrinkles). It gives it a
            recognizable, characteristic flavor that was absent from Ebisédian where
            grammar dominates over prosody. Recently, I even discovered that TF
            prosody appears to be based on an alternating stressed/unstressed scheme
            that starts counting from the *end* of a clause, which causes certain
            clause structures to prefer a certain order to NPs so that key NPs
            receive emphasis over peripheral NPs. That was a very happy discovery;
            it gives a naturalistic ordering of utterances rather than just a
            generic anything-goes "it can be in any order 'cos the grammar says so"
            deal. Even natlangs with "free" word order, like Russian, seems to
            gravitate toward a certain default order that gives it a characteristic
            flavor. Switching that default order (which is allowed by grammar) tends
            to give it an overt change in emphasis that can convey drama, irony,
            etc..


            > There are some decidedly non-IE adverbs, though, among which are what
            > I call "adverbs of manner". These are verb modifiers that serve that
            > function of what in English requires an auxiliary verb construction;
            > for example:
            >
            >     tapa    ha         buara   na!
            >     walk(V) start(ADV) volcano RCP:MASC
            >     Start walking to the volcano!
            >
            > In the English translation, the main verb is "start" modified by
            > "walking"; in TF, on the contrary, the main verb is "walk" and "start"
            > is an adverb ("startingly"?). In the same vein:
            >
            >     tapa    bat       buara   na!
            >     walk(V) stop(ADV) volcano RCP:MASC
            >     Stop walking to the volcano!
            >
            > Quite a good number of constructions in English that requires an
            > auxiliary verb are subsumed by adverbs-of-manner in TF:
            >
            >     arap       pera     beira sei.
            >     pick_up(V) try(ADV) stone CVY:FEM
            >     Try to pick up the stone!
            >
            >     tsana   irei          maharan so.
            >     talk(V) continue(ADV) story   CVY:NEUT
            >     Continue to tell the story.
            >
            > RM but what does TF do when the "aux." is the main verb, as in an imperative:
            >  Start now! Don't walk there! Do go on (continue)!  etc.???

            The examples above are all imperatives, in fact. It seems impossible to
            say something like "try!" without also specifying the action one is
            supposed to try. So there's no (direct) analogue to English expressions
            like "Can you at least try?" or "please stop!". In TF, you have to
            specify what to try, and what needs to stop, since "try" and "stop" are
            adverbs and can't stand on their own.

            TF imperatives are indicated by verb-initial word order (and often
            omitted finalizer), whereas indicatives are indicated by verb-second
            order. So, for comparison, you have:

            tara' ka tsana irei maharan so
            3SG ORG:MASC talk(V) continue(ADV) story CVY:NEUT
            He continues telling the story.

            tsana irei maharan so aniin
            talk(V) continue(ADV) story CVY:NEUT FIN
            Continue telling the story!

            You can also make hortatives by including the 3SG pronoun in imperative
            word order:

            tsana irei tara' ka maharan so
            talk(V) continue(ADV) 3SG ORG:MASC story CVY:NEUT
            Let him continue telling the story!
            (Not in the sense of "allow him", but in the sense of a
            hortative, "he shall continue to tell the story".)

            Finalizers are optional in imperatives, since an imperative may or may
            not be obeyed, so the action may or may not be completed; but can be
            included to indicate forcefulness (emphasize that the request action
            *must* be completed).

            tsana irei maharan so aniin!
            talk(V) continue(ADV) story CVY:NEUT FIN
            Continue telling the story [to the end]!


            > They form an interesting class, because they exhibit dialectal
            > differences in TF. The san faran are divided into the northern tribes
            > and the southern tribes, and there is a prominent volcano that roughly
            > forms the boundary between north and south. So you have the adverb
            > _buaranui_ (from _buara_ "volcano" + -n- [genitive?  ancient
            > nominalizer?  something else?] + -ui), which to a Northerner means
            > "southwards", but to a Southerner means "northwards". :)
            >
            > RM this happens in a lot of Austronesian lgs. as well, where
            > "sea(wards)" can mean east, west, north, south depending on the
            > island/region; same with "inland"; and others

            Cool! Yet another Austronesian feature in TF. :-)


            > What do y'all do when you rediscover old conlang notes that had been
            > forgotten for many years, and find that it contradicts what you had
            > subsequently developed in your conlang?
            >
            > RM I shake my head and go, whatever made me do that??? :-))))

            Hehe... I guess in the case of that TF poem, I *could* potentially go
            back and retcon it to conform to my updated vision of proto-TF. :)
            That's the nice thing about conlanging: you get to work with
            2-dimensional time (in-world time and out-of-world time, which need not
            correspond linearly).


            > Except in the case of Gwr, when I discovered several original forms
            > that didnt comply with the rules I eventually formulated...........In
            > that case, don't repaid, create anew....

            Or you could leave them as irreducibly irregular forms inherited from
            who-knows-when. Like English strong verbs that still retain PIE ablaut,
            or calcified expressions that no longer follow contemporary grammar
            rules.


            On Mon, Apr 22, 2013 at 07:41:04AM -0700, Roger Mills wrote:
            > Afterthought:
            >
            > Some of the "finalizers" look a bit like old serial verb
            > constructions.

            Well, that could be an artifact of my poor glossing. :-P

            But, it's an interesting thought. What if proto-TF actually had serial
            verb constructions? It will have some interesting consequences, for
            sure. Hmmm... I'll have to think about that a bit more!


            T

            --
            Computerese Irregular Verb Conjugation: I have preferences. You have
            biases. He/She has prejudices. -- Gene Wirchenko
          Your message has been successfully submitted and would be delivered to recipients shortly.