A Braille Orthography for tlhIngan—Stovepipe Wells-Jensen SpecGram Vol CLII, No 2 Contents <i>The Other Sino-Tibetan Languages</i>—Book Announcement from Psammeticus Press

Systematic Suppletion: An Investigation of Ksotre Case Marking

By Lawrence R. Muddybanks, Ph.D.


Little research has been done on the Ksotre language of
Tha mo bhàta-foluaimein
loma-làn easgannan
— Scots Gaelic
northern Lithuania, and that which has been done has been rather unenlightening.1 The present paper aims to not only expand the body of research on the Ksotre language, but also to introduce a phenomenon found, thus far, in no other natural language on Earth. Without further ado, then, I present the case marking system of Ksotre.

Case in Ksotre

Ksotre has fourteen cases and three numbers, as shown below:

Furthermore, case and number marking is rather straightforward. There is a mixture of suffixes and prefixes which account for each of the case and number features. They are summarized in the table below:

  1. Number:
    • Singular: -ø
    • Dual: -aw
    • Plural: -e

  2. Grammatical Cases:
    • Nominative: -ø
    • Accusative: -ir
    • Partitive: -et
    • Genitive: -uz
    • Dative: -ort
    • Instrumental: -ag
  1. Local Cases:
    • Adessive: k(e)-
    • Allative: kr(e)-
    • Ablative: ks(e)-
    • Inessive: p(i)-
    • Illative: pr(i)-
    • Elative: ps(i)-
    • Translative: sr(u)-
    • Prolative: l(a)-
Vehicolul meu pe pernă
de aer e plin cu ţipari
— Romanian

Though it may be considered a typological oddity by some that an overt number suffix will follow an overt case suffix, the true mystery of the Ksotre case system lies in the fully inflected form of a given noun. Below is a common example, using the noun lefked, “girl”:

lefked, “girl” Singular Dual Plural
Nominative lefked lefkedaw lefkede
Accusative lefkedir lefkediraw lefkedire
Partitive lefkedet lefkedetaw lefkedete
Genitive lefkeduz lefkeduzaw lefkeduze
Dative lefkedort lefkedortaw lefkedorte
Instrumental lefkedag lefkedagaw lefkedage
Adessive klefked klefkedaw klefkede
lefked, “girl” Singular Dual Plural
Allative krelefked krelefkedaw krelefkede
Ablative kselefked kselefkedaw kselefkede
Inessive plefked plefkedaw plefkede
Illative prilefked prilefkedaw prilefkede
Elative psilefked hontordarus psilefkede
Translative srulefked srulefkedaw srulefkede
Prolative lalefked lalefkedaw lalefkede

The paradigm in (2) at first appears rather uninteresting, in that most cells are filled with fully inflected forms that one would expect, given the case and number affix
El meu aerodesllissador
està ple d’anguiles
— Catalan
inventory in (1). Looking at the elative dual cell, however, one finds hontordarus, which not only has neither the dual suffix -aw nor the elative prefix ps(i)-, but also appears to bear no resemblance to the stem lefked whatsoever. This is, in fact, a case of suppletion, not unlike go/went in English, or ir/va/fue in Spanish. Were that the end of it, the only interest of this paradigm would be the fact that the suppletion occurs in a rather unexpected place (i.e., one doesn’t usually expect to find a second stem used exclusively in the elative dual). Consider, however, the random sampling of native partially declined Ksotre nouns below:

Noun Acc. Sg. Dative Dual Gen. Pl. Adessive Sg. Inessive Dual Elative Dual
ertrav, “bee” ertravir ertravortaw ertravuze kertrav prertravaw wit
topos, “kettle” toposir toposortaw toposuze ketopos pritoposaw ilili
tsuip, “pen cap” tsuipir tsuiportaw tsuipuze ketsuip pritsuipaw orso
ankuf, “legislature” ankufir ankufortaw ankufuze kankuf prankufaw deskovurtiva
saval, “student” savalir savalortaw savaluze kesaval prisavalaw a
elalit, “fire” elalitir elalitortaw elalituze kelalit prelalitaw srostra
ovult, “indecency” ovultir ovultortaw ovultuze kovult provultaw kepim
mim, “mother” mimir mimortaw mimuze kemim primimaw psertravaw

After having elicited over two thousand native Ksotre nouns which follow the same pattern, I’m afraid I have no choice but to state that the elative dual form of a given noun in Ksotre will be suppletive. In this way, Ksotre displays a kind of systematic suppletion that no linguist could ever have predicted would exist in any possible natural language.

— Cantonese


Presented with data such as that in (3), one is left with a number of questions. First, is it true that the forms are, in fact, suppletive? A quick glance at the elative dual column in (3) should suggest that, indeed, the forms are suppletive. Though, for example, both etrav and wit contain a t, and saval and a contain an a, one would be hard-pressed to come up with a set of transformational rules or OT constraints which would produce the precise elative dual form for each lexeme.2 One could, I suppose, come up with a word and paradigm analysis, à la Bochner, that would look something like this:

Z, Nom. Sg.
Z, Acc. Sg.
Z, Dat. Plu.
Z, Ela. Dual

The rule above states simply that any noun of the form X associated with the nominative singular and some meaning Z (e.g., “bee”) will be in a relationship with other forms, such that Xir is the accusative singular, Xorte is the dative plural, etc., and that, given these relations, one can expect an unpredictable suppletive form Y in the elative dual. Of course, the very fact that this analysis is attractive is proof enough that the analysis is wrong, and that
Moja lebdjelica
je puna jegulja
— Croatian
lexicalist theories of grammar are misguided. Therefore, the presentation of a real analysis shall have to wait. Let’s move on.

The second question one is forced to ask is, even though it would be utterly bizarre to have an empty cell in a paradigm, is it the case that, perhaps, the elative dual forms are simply separate lexemes, as unrelated as “car” and “pickle”? Grammatical analysis and common sense would suggest otherwise, but the fact remains that the suppletive forms just look drastically different. To answer this question, let me present a list of partially declined loan words:

Noun Acc. Sg. Dative Dual Gen. Pl.
televizdep, “television” televizdepir televizdeportaw televizdepuze
intanedep, “internet” intanedepir intanedeportaw intanedepuze
profodep, “balloon” profodepir profodeportaw profodepuze
lingwizdep, “lunatic” lingwizdepir lingwizdeportaw lingwizdepuze
Noun Adessive Sg. Inessive Dual Elative Dual
televizdep, “television” ketelevizdep pritelevizdepaw psitelevizdepaw
intanedep, “internet” kintanedep printanedepaw psintanedepaw
profodep, “balloon” keprofodep priprofodepaw psiprofodepaw
lingwizdep, “lunatic” klingwizdep prilingwizdepaw psilingwizdepaw

As you can see, the elative dual forms work exactly as one would expect, given the information in (1), but work exactly the way one would not expect given the sample paradigms in (2) and (3).3 The fact, though, that the morphology is available to loan words suggests that the elative dual is, indeed, a cell
Mit luftpudefartøj
er fyldt med ål
— Danish
in the standard nominal paradigm, and that, consequently, the elative dual forms for native Ksotre nouns are not separate lexemes, but, rather, count as a case of systematic suppletion.

To conclude, looking at such a system as this one, one has to wonder: will it survive? After all, in everyday discourse, how often does one say that something or someone comes out of two of something? As it turns out, quite a bit. Having done a corpus analysis of the local Ksotre newspaper archives, I present to you the likelihood that a given case and number combination will be used with a random noun:

lefked, “girl” Singular Dual Plural Total % (Case)
Nominative 0.2% 1.3% 0.1% 1.6%
Accusative 0.4% 2.2% 0.1% 2.7%
Partitive 0.6% 3.7% 2.1% 6.4%
Genitive 0.6% 0.7% 4.8% 6.1%
Dative 0.9% 0.4% 1.1% 2.4%
Instrumental 0.9% 2.7% 0.1% 3.7%
Adessive 1.1% 2.3% 1.1% 4.5%
Allative 0.4% 2.5% 0.3% 3.2%
Ablative 0.4% 4.5% 0.8% 5.7%
Inessive 0.7% 2.5% 1.3% 4.5%
Illative 0.3% 1.9% 0.7% 2.9%
Elative 0.1% 48.9% 0.1% 49.1%
Translative 1.2% 5.5% 0.5% 7.2%
Prolative4 0% 0% 0% 0%
Total % (Number) 7.8% 79.1% 13.1% 100%

As you
Puno ng palos ang
aking hoberkrap
— Tagalog
can see, the elative dual is the most common case/number combination used (48.9%). Further, the dual number (79.1%) and the elative case (49.1%) are the most common number and case, respectively. If the Ksotre newspaper archives can be considered a fair representation of the Ksotre language, then it would appear that the elative dual is quite frequent in everyday Ksotre. My personal experience with Ksotre speakers has served only to reinforce this hypothesis. Consider the most common Ksotre greeting:

7.   Sru-kevdeg-aw v-e-gors-efska istop-ir-aw plimbu kre-numan-ilt-aw!5
/Trans.-cat-Dual 3rd.Dual.Sbj.-3rd.Dual.Obj.-eat-Optative rooster-Acc.-Dual bowl.Ela.Dual All.-daughter-2nd.Dual.Poss.-Dual!/
“Greetings!” (Literally, “May two cats eat two roosters out of two bowls for you two’s two daughters!”)

Indeed, one may not order a meal, ask for directions, cheer at a football match, or use the restroom without having a full command of all elative and dual forms of each noun in Ksotre. Some early acquisition evidence suggests that children learn the elative dual form of a noun often before they learn that noun’s nominative singular, accusative singular or ablative plural forms.6 So despite the peculiar nature of the suppletive elative dual forms, they appear to be here to stay.

Hoverkraftimin ici
yılan balıǧı dolu
— Turkish


Little has been concluded about the bizarre nature of the elative dual in Ksotre. Ima Morontz (ms) theorizes in a yet-to-be-published paper that perhaps the aberrant elative dual forms are, in fact, borrowings into Ksotre.7 As evidence, he cites the declension of the gender-neutral third person pronoun rof (partial paradigm shown below):

Noun Acc. Sg. Dative Dual Gen. Pl. Adessive Sg. Inessive Dual Elative Dual
rof, “s/he” rofir rofortaw rofuze krof profaw sunnus

The elative dual form sunnus, Morontz notes, is identical in form to the elative dual declension of the gender-neutral third person pronoun (sån) of the Finno-Ugric language Lule Sámi.8 He contends that the Ksotre language takes part in a
माझी होडी मासळयांनी
भरली आहे
— Marathi
kind of linguistic exchange program, borrowing elative dual forms from various languages, and, in turn, sending their native elative dual forms in exchange. Though I hesitate at this time to propose an alternate theory, I’d like to suggest that the metaphor utilized by Morontz, at least, is flawed, in that the Lule Sámi language is neither missing its elative dual, nor employing the “native” Ksotre form psirofaw. Additionally, a search for other languages from which the Ksotre have borrowed elative dual forms has, thus far, proved fruitless.9

In the opinion of the author, further research is required into the nature of the curious case system of Ksotre before anything definitive can be said. It is my hope, however, that the data presented above will whet the appetites of morphologists worldwide, and give birth to new research programmes centering on this little-studied, yet wonderfully rich and perplexing language.

1 See, for example, Snodgrass (2001), Ksotre Phonology: It has one.

2 The elative dual form of mim,
Mi aerodeslizador está
lleno de anguilas
— Spanish
“mother”, is especially vexing, as it’s identical to what one would expect to be the elative dual form of ertrav, “bee”. This, apparently, is entirely coincidental. I tore my hair out (quite literally) for weeks simply trying to get a speaker of Ksotre to admit that the elative dual of mim looks as if it should be the elative dual of ertrav. Unfortunately, none of them had the faintest idea what I was talking about. One, in fact, took great offense, believing that I had suggested that his mother was born of two bees. One black eye and two bruised ribs later, I gave up.

3 The word televizdep, it should be noted, may be undergoing a process of cultural assimilation, as it were. Of the nine hundred or so instances of televizdep in my corpus, almost two hundred occur without the -dep suffix. What’s perhaps most interesting is that of the over four hundred instances of the elative dual form of televizdep, there are more than ninety occurrences of a native-like elative dual form, ripusfel, suggesting that, perhaps, televizdep will soon be just like any other native noun of Ksotre.

4 I couldn’t find a single occurrence of the prolative in any number. I would not even have known that it existed had I not encountered an elderly monk who, as he died in my arms, whispered to me the one verb with which the prolative is used. That verb, however, is so vulgar
Իմ սավառնաթիռում
լիքը օձաձուկ է
— Armenian
that I have never heard any other speaker of Ksotre use it, and have had to flee for my life on more than three occasions after asking a consultant if they’d ever heard of the verb in question.

5 The nominative singular form of “rooster” in Ksotre is rustur.

6 See Pearl and Stücke’s groundbreaking article, Children and Language: Two things that don’t mix (2005).

7 See Morontz (to appear), Why I’m Better Than You at Linguistics: New evidence from Ksotre.

8 For the curious, a partial paradigm is listed below:
  Nom. Gen. Inessive Elative
Singular sån suv sujna sujsta
Dual såj sunnu sunnun sunnus
Plural sij siljá sijan sijas
എന്റെ പറക്കും-പേടകം
നിറയെ വ്ളാങ്കുകളാണു
— Malayalam
9 There is also historical evidence which suggests that the form sunnus derived from an earlier attested form *somnus via a series of sound changes which affected the entire language. Of course, as this evidence is diachronic, it has no place in modern linguistics.

A Braille Orthography for tlhIngan—Stovepipe Wells-Jensen
The Other Sino-Tibetan Languages—Book Announcement from Psammeticus Press
SpecGram Vol CLII, No 2 Contents