{"id":2256,"date":"2024-12-27T14:13:25","date_gmt":"2024-12-27T19:13:25","guid":{"rendered":"https:\/\/www.wellformedness.com\/blog\/?p=2256"},"modified":"2025-01-01T14:24:36","modified_gmt":"2025-01-01T19:24:36","slug":"languages-two-is","status":"publish","type":"post","link":"https:\/\/www.wellformedness.com\/blog\/languages-two-is\/","title":{"rendered":"Languages with two i’s"},"content":{"rendered":"

[I’ve edited this after the fact to add a new case study. -KBG.]<\/strong><\/p>\r\n

Doing some research on underspecification and its use as an exceptionality-generating device for our work on Logical Phonology, I have encountered an common pattern: there are quite a few languages, of wide genetic and geographic distribution, whose description suggests their phonological inventory contains two underlying segments ordinarily realized as [i(\u02d0)] but showing different phonological behaviors. The following is a brief catalog followed by some discussion.<\/p>\r\n

Barrow Inupiaq<\/h1>\r\n

I have already discussed this well-known case<\/a>, which is discussed by Dresher (2009:\u00a77.2.1), who in turn has the data from Kaplan (1981:\u00a73.22). In a number of dialects of Eskimo-Aleut, a four-phoneme vowel system *i<\/em>, *u<\/em>, *\u0259<\/em>, *a<\/em> has been simplified to an ordinary three-vowel system by the merger of\u00a0*\u0259<\/em>\u00a0into *i<\/em>. However, in Barrow Inupiaq, there is a “strong\u00a0i<\/em>” (<\u00a0*i<\/em>) which triggers palatalization of a following coronal consonant whereas “weak\u00a0i<\/em>” (< *\u0259<\/em>) does not.\u00a0<\/p>\r\n

(1) Barrow Inupiaq (Kaplan 1981:\u00a73.22, his 27-29):<\/p>\r\n

a. iglu \u2018house\u2019, iglul<\/strong>u \u2018and a house\u2019, iglun<\/strong>ik \u2018houses\u2019
b. ini \u2018place\u2019, inil<\/strong>u \u2018and a place\u2019, inin<\/strong>ik \u2018places\u2019
c.\u00a0iki \u2018wound\u2019, iki\u028e<\/strong>u \u2018and a wound\u2019, iki\u0272<\/strong>ik \u2018wounds\u2019<\/p>\r\n

Presumably, the stem-final\u00a0i\u00a0<\/em>in (1b) is weak and the one in (1c) is strong. One possible way to derive this (slightly different from what I said in my earlier post) is to treat palatalization as triggered by [\u2212Back, +High] strong i<\/em>, leave weak\u00a0i<\/em> underspecified for Back, and then use the following late redundancy rule to fill in the missing specification for weak\u00a0i.<\/em><\/p>\r\n

(2) [+High] \u2294 {\u2212Back}<\/p>\r\n

Given the definition of unification, this will non-vacuously apply to weak\u00a0i,<\/em>\u00a0vacuously apply to strong i<\/em>, and unification with [+High, +Back] \/u\/ will fail; \/a\/ does not meet the structural description because it is [\u2212High].<\/p>\r\n

Blackfoot<\/h1>\r\n

Charles Reiss (personal communication) brings my attention to a case in Blackfoot. Frantz (2017:\u00a76B) describes the situation clearly:<\/p>\r\n

\r\n

The initial vowel of stem ipii<\/em> \u2018enter,\u2019 unlike the initial vowel of itsiniki<\/em> \u2018tell a story,\u2019\u00a0always causes a preceding k<\/em> to be replaced by the affricate ks<\/em>. We will speak of this phenomenon as breaking<\/strong> of k<\/em>, and of the i<\/em> which is involved\u00a0as a breaking i<\/em><\/b>. For any morpheme which begins with i<\/em> we need to know\u00a0whether that i<\/em> is a breaking i<\/em> or not; if it is a breaking i<\/em>, then if it immediately follows a morpheme ending in k<\/em> we know that the k<\/em> will be replaced by ks<\/em>.<\/p>\r\n<\/blockquote>\r\n

Naturally enough, Frantz proposes that there are two\u00a0i<\/em>‘s in Blackfoot. Thus far, one would expect the analysis to be parallel to our analysis of palatalization in Barrow Inupiaq, with “catalytic” (breaking), fully-specified \/i\/ and “quiescent” (non-breaking) underspecified segment (which Frantz writes as \/I\/). But there is there is one wrinkle; a few pages later Frantz (op. cit.:\u00a76D) reports that the second person prefix\u00a0k-\u00a0<\/em>is “impervious to breaking”. Thus, it seems like we also need two k<\/em>‘s: one “mutable” (breaking), underspecified segment we can write as \/K\/, and one “inalterable” (non-breaking), fully-specified \/k\/. Breaking then occurs only in the sequence \/…K-i…\/.<\/p>\r\n

Czech<\/h1>\r\n

Anderson & Browne (1973; henceforth AB) discuss several vocalic alternations in two registers of Czech. I will focus on the literary language (“LL” for AB; henceforth just Czech) and put aside additional complications associated with their analysis of the common spoken register, which makes some possibly problematic simplifying assumptions for sake of argument.<\/p>\r\n

The surface front vowels of Czech are [i, i\u02d0, \u025b, \u025b\u02d0].1<\/sup> However, specific instances of these three vowels either trigger or do not trigger palatalization of the preceding consonant.2<\/sup> For example, the masculine animate nom.pl. -i<\/em> [-i] is palatalizing, but the feminine nom.pl -y<\/em> [-i] and the gen.pl. -ych <\/em>[-ix]\u00a0are not:<\/p>\r\n

(3) sest\u0159in [s\u025bstr\u031din<\/strong>] ‘sister’s’, sest\u0159ini\u00a0[s\u025bstr\u031di\u0272<\/strong>i] masc.anim. nom.pl.,\u00a0
sest\u0159iny\u00a0[s\u025bstr\u031din<\/strong>i] fem. nom.pl., sest\u0159inych [s\u025bstr\u031din<\/strong>ix] gen.pl.<\/p>\r\n

The same is true of [\u025b]-initial suffixes. For example, the -\u011b<\/em> [-\u025b] prep.sg. suffix is palatalizating, but the -ech <\/em>[-\u025bx] prep.pl. is not (note the -o\u00a0<\/em>suffix in the citation form is the neut. nom.sg.):<\/p>\r\n

(4) okno [okn<\/strong>o] ‘window’, okn\u011b [ok\u0272<\/strong>\u025b] prep.sg., oknech [okn<\/strong>\u025bx] prep.pl.<\/p>\r\n

Czech [i, i\u02d0] are reflexes of both Old Czech i<\/em>(\u02d0<\/em>)\u00a0and \u0268<\/em>(\u02d0<\/em>), the latter two merging into the former two respectively. This distinction is preserved in the writing system as [i, i\u02d0<\/em>]\u00a0from earlier \u0268<\/em>(\u02d0<\/em>) is written <y, \u00fd> respectively. AB adopt the sensible proposal that this distinction is also preserved phonologically. That is, they suggest non-palatalizing <y, \u00fd> are underlyingly [+High, \u2212Low, \u2212Round, +Back] and palatalizing <i, \u00ed> are [+High, \u2212Low, \u2212Round, \u2212Back]. While they do not sketch out a full analysis, presumably <y, \u00fd> are fronted after palatalization has applied. Alternatively (and more in tune with the approach of Gorman and Reiss 2024), one might instead propose that <y, \u00fd> are [+High, \u2212Low, \u2212Round] and underspecified for [Back]. Palatalization will be triggered by a following [\u2212Back] vowel, and then the following unification rule will (non-vacuously) front <y, \u00fd>:<\/p>\r\n

(5) [\u2212Low, \u2212Round] \u2294 {\u2212Back}<\/p>\r\n

As written, this will also also handle the parallel behavior of non-palatalizing <\u011b>, which gives it the same featural representation as palatalizing <e> [\u025b]. Under either account, though, Czech has two i<\/em>‘s (in long and short variants) and\u00a0two e<\/em>‘s too.<\/p>\r\n

Hungarian<\/h1>\r\n

Charles Reiss (p.c.) draws my attention to the existence of a roughly 60\u00a0 Hungarian noun stems (Sipt\u00e1r & T\u00f6rkenczy 2000:\u00a73.2.2; henceforth ST) which are apparent exceptions to harmony in that they have a front vowel but select for back-vowel allomorphs. For instance, “regular” v\u00edz\u00a0<\/em>‘water’ has the harmonic dative v\u00edzne<\/strong>k\u00a0<\/em>and ablative v\u00ed<\/strong>zt\u0151<\/strong>l<\/em>, and so on. But there are some exceptions, which ST call “antiharmonic”.<\/p>\r\n

(6) Antiharmonic stems (after ST:68, their 33):
a. h\u00edd ‘bridge’, h\u00ed<\/strong>dna<\/b>k dat.sg. (*h\u00eddne<\/strong>k), h\u00eddt\u00f3<\/strong>l\u00a0abl.sg. (*h\u00ed<\/strong>dt\u0151<\/strong>l)
b.\u00a0szi<\/strong>d\u00a0‘scold’, szi<\/strong>dha<\/strong>t ‘may scold’, szi<\/strong>d\u00f3 <\/strong>‘scolding’
c. h\u00e9j ‘crust’, h\u00e9<\/strong>jna<\/strong>k dat.sg., h\u00e9<\/strong>jt\u00f3<\/strong>l abl.sg.<\/p>\r\n

According to ST, most antiharmonic stems are in i<\/em> or \u00ed<\/em> as in (6ab), and just a few are in \u00e9<\/em> as in (6c). While these are traditionally understood as lexical (morpheme-level) exceptions, one possible account is to treat antiharmonic \u00ed<\/em>, i<\/em>, or\u00a0\u00e9 <\/em>as having a different underlying specification than harmonic \u00ed<\/em>, i<\/em>, and \u00e9<\/em>.<\/p>\r\n

There are a\u00a0 number of possible ways this might work out. According to ST, \u00ed<\/em>, i<\/em>, and \u00e9 <\/em>are not so much “harmonic” but rather “neutral”. They claim the front-harmonic suffix allomorphs (like –ne<\/strong>k\u00a0<\/em>and –t\u0151l<\/em>) are the default allomorphs. Indeed, stems consisting of a non-neutral vowel like \u00e1\u00a0<\/em>[a\u02d0],\u00a0a\u00a0<\/em>[\u0252],\u00a0o\u00a0<\/em>[o] or \u00f3\u00a0<\/em>[o\u02d0] followed by \u00ed<\/em>, i<\/em>, and \u00e9\u00a0<\/em>select the back allomorphs, and do so without exception<\/span>, suggesting \u00ed<\/em>, i<\/em>, and \u00e9<\/em> are simply transparent to harmony.<\/p>\r\n

(6) Back-neutral stems (after ST: loc. cit<\/em>., their 34b):
a. pap\u00edr ‘paper’, pa<\/strong>p\u00ed<\/strong>rna<\/strong>k dat.sg., pa<\/strong>p\u00ed<\/strong>rt\u00f3<\/strong>l abl.sg.
b. d\u00f3zis ‘dose’, d\u00f3<\/strong>zi<\/strong>sna<\/strong>k dat.sg., d\u00f3<\/strong>zi<\/strong>st\u00f3<\/strong>l abl.sg.
c. k\u00e1v\u00e9 ‘coffee’, k\u00e1<\/strong>v\u00e9<\/strong>na<\/strong>k dat.sg., k\u00e1<\/strong>v\u00e9<\/strong>t\u00f3<\/strong>l abl.sg.<\/p>\r\n

If we adopt ST’s perspective, then one possibility is that harmonic-neutral \u00ed<\/em>, i<\/em>, and \u00e9\u00a0<\/em>are specified\u00a0 [\u2212Low, \u2212Round] and antiharmonic vowels are [\u2212Low, +Back, \u2212Round]\u2014keeping them distinct from underlying u\u00a0<\/em>[u], \u00fa<\/em> [u\u02d0], and \u00f3 [o\u02d0]\u2014and become [\u2212Back] later in the derivation. The following context-free rules, applied in order after harmony has been computed, ought to suffice.3<\/sup><\/p>\r\n

(7) [\u2212Low, \u2212Round] \u2216 {+Back}
(8) [-Low, -Round] \u2294 {\u2212Back}<\/p>\r\n

(7) gives harmony-neutral and antiharmonic front vowels the same representation, applying non-vacuously only to the latter,\u00a0and (8) insures that both are realized as [\u2212Back] on the surface. The [-Round] condition is essential to exempt front round vowels\u2014namely \u00fc\u00a0<\/em>[y], \u0171<\/em> [y\u02d0],\u00a0\u00f6<\/em> [\u00f8], and \u0151<\/em> [\u00f8\u02d0]\u2014and the [-Low] condition exempts low vowels, which are back-harmonic.<\/p>\r\n

Kashaya<\/h1>\r\n

Buckley (1994) describes four unusual rules in Kashaya, given below in extensional (i.e., non-featural) notation.<\/p>\r\n

(9) i -> a \/ m _
(10) i -> u \/ d _
(11) V -> a \/ q _
(12) V -> o \/ qw<\/sup> _<\/p>\r\n

While (9-10) appear to be quite regular processes, he notes that 3 out of 21 i-<\/em>initial suffixes (the inchoative\u00a0-ibic\u00a0<\/em>and the reflexives -iyic\u2019<\/em>\u00a0and -ic\u2019<\/em>) fail to undergo them. Similarly, most\u00a0i<\/em>‘s undergo (11-12) but not the previously mentioned i<\/em>-initial suffixes. With (11) the result is […ki…] instead of *[…qa…]; with (12) the result is […qo…] instead of *[…qw<\/sup>o…].<\/p>\r\n

Buckley proposes that Kashaya has two i<\/em>‘s\u2014mutable \/i\/ and inalterable \/\u00ee\/\u2014which are surface-identical except with respect to (9-12). He proposes that \/\u00ee\/ is [+High] whereas \/i\/ is underlyingly underspecified for it. This analysis can be easily translated into Logical Phonology. For instance, (9) can be respecified as follows.4<\/sup><\/p>\r\n

(13)\u00a0 [+Vocalic] \u2294 {-High, +Low} \/ [+Labial, +Nasal, …] _<\/p>\r\n

(13) applies non-vacuously<\/span> to only one segment: underspecified \/i\/. Both High and Low must be specified since Kashaya also has mid vowels \/e, o\/. Then a context-free redundancy rule can then be used to “merge” \/i, \u00ee\/, non-vacuously applying to any underspecified \/i\/ remaining after (13) and the Logical Phonology analogues of (10-12).<\/p>\r\n

(14) [+Vocalic] \u2294 {+High, -Low}<\/p>\r\n

Spanish<\/h1>\r\n

The final case concerns the so-called raising verbs of the Spanish 3rd conjugation. This is a well-known case and you can read my<\/del> our take on it in Gorman & Reiss 2024 (\u00a74) here<\/a>. Briefly, we argue that the “raising” verbs have an underlying vowel underspecified for High which is realized as\u00a0i\u00a0<\/em>except when it is followed by a [-Back, +High] vowel in the following syllable. Verbs with the underspecified i<\/em>-like vowel, like the pe<\/strong>dir-pi<\/strong>do<\/em> ‘ask for (s.t.)’ are thus distinct from verb stems with non-alternating i\u00a0<\/em>(e.g., vi<\/strong>vir-vi<\/strong>vo ‘<\/em>live’) or e\u00a0<\/em>(sume<\/strong>rgir<\/em>–sume<\/strong>rjo<\/em> ‘submerge’).<\/p>\r\n

Discussion<\/h1>\r\n

Summarizing our analyses:<\/p>\r\n