Proto-Isthmus

Proto-Isthmus
Period	-3000 YP
Spoken in	Huyfárah/Isthmus region
Total speakers	unknown
Writing system	none
Classification	Eigə-Isthmus languages  Proto-Isthmus
Typology
Basic word order	SOV?
Morphology	fusional
Alignment	NOM-ACC
Credits
Created by	Radius, Corumayas, Cedh

Proto-Isthmus was spoken perhaps a thousand years before Ndak Ta (ca. -3000 YP), in the isthmus region separating the main continent of Peilaš from the northeastern subcontinent of Siixtaguna. The Isthmus languages descended from it form one of the two divisions of the Eigə-Isthmus family; the other division is the Eigə Valley languages, comprising Ngauro, Meshi, and the Miwan languages.

Seven Isthmus languages are attested by the first millenium YP; they fall into two branches. In Western Isthmus we find the closely related sister languages Feråjin and Faraghin, spoken in Huyfárah, and the more divergent Boésin (Fáralo exonym), spoken in Qedik territory north of the Šišin Mountains. Three Eastern Isthmus languages survive on the north and northwest coasts of the Siixtaguna subcontinent (Kennan, Kietek, and Ka'alikora), and a fourth remains on the Isthmus itself, known in Fáralo as Doroh. Of them all, only Doroh and Kennan were spoken by a great number of people, perhaps half a million apiece in 100 YP, and only these and perhaps Feråjin (or rather, descendants of them) survived into the second millennium YP.

Diachronics of Proto-Isthmus

NOTE: not all diachronic changes between Proto-Isthmus and its daughters are listed, only those that are known at this time. However, the list of developments from Proto-Isthmus to Faraghin is more or less complete.

Phonemes that have been reconstructed for Proto-Isthmus are /p t ʈ ts k b d ɖ dz g f s m n l j a e i o u/.

Prefixes for pronouns included:

• as- (genitive/accusative)
• dza- (dative)

We will be using the pronouns of Isthmus languages to illustrate their development. The reconstructed personal pronouns of Proto-Isthmus are as follows.

	NOM	GEN/ACC	DAT
1sg	da	asda	dzada
1pl	guʈ	asguʈ	?
2sg	tujn	astujn	dzatujn
2pl	fe	asfe	?
3sg-anim	njo	asnjo	dzanjo
3sg-inan	mis	asmis	?
3pl	ludz	asludz	?
int/rel-anim	kej	askej	?
int/rel-inan	bej	asbej	?

Though we can posit the forms of the other dative pronouns, shown as "?" (e.g. **dzaguʈ), no reflexes of such forms can be identified in any Isthmus language. Indeed, Faraghin is the only Isthmus language to retain any of the dative pronouns, and their reconstruction - or if they were truly even datives - remains uncertain. The only corroboration comes from, firstly, certain locative adverbs in Doroh that appear to be reflexes of Proto-Isthmus nouns with the same morphology; and secondly, some adverbial noun forms that appear in archaic Boésin poetry, in which the meaning is directional (comparable to English -wards).

Western Isthmus

Pre-Western Isthmus

Developments leading to pre-Western Isthmus include the following.

• First, stress moved to the first syllable of each word that had an onset of at least one consonant. Then, any unstressed vowels before the onset were lost. At some point after this, some final vowels were also dropped, particularly from grammatical words.

• s became voiced when adjacent to a voiced consonant.

• j was lost as a phone in all environments. At the beginning of words and between vowels, it merged into dz. It was also lost outright when adjacent to i, except that the sequences *tji *dji merged into tsi dzi, and medial *kji *gji became ksi gzi. After this, the sequences oj, ej, and aj merged into the diphthong ai, after which all remaining instances of j before or after a vowel merged into ɨ.

At this stage, the inventory of personal pronouns is reconstructed as follows.

	NOM	GEN/ACC	DAT
1sg	da	zda	dzad
1pl	guʈ	zguʈ
2sg	tɨn	stɨn	dzatɨn
2pl	fe	sfe
3sg-anim	nɨ	znɨ	dzan
3sg-inan	mis	zmis
3pl	ludz	zludz
int/rel-anim	kai	skai
int/rel-inan	bai	zbai

Western Isthmus

Further developments led to Western Isthmus, the last common ancestor of Boésin and the Ferogh languages.

• Before or after s, or z, all stops became their corresponding fricatives:
  p t ʈ k b d ɖ g > ɸ θ ʂ x β ð ʐ ɣ / _s, _z, s_, z_
• This included the affricates ts and dz. Then, s and z were lost not only from clusters with the new fricatives but also from all other clusters, no matter their position in the cluster. This too included the former affricates.

• The fricative ɸ subsequently merged into the preexisting phoneme f. The retroflex fricatives then became postalveolar ʃ ʒ, and the retroflex stops became postalveolar affricates tʃ dʒ.

• Sporadically, obstruents at the same POA in adjacent syllables dissimilate (e.g. dadaða- > daðada- "dance").

The reconstructed pronoun table for Western Isthmus follows.

	NOM	GEN/ACC	DAT
1sg	da	ða	ðad
1pl	gutʃ	ɣutʃ
2sg	tɨn	θɨn	ðatɨn
2pl	fe	fe
3sg-anim	nɨ	nɨ	ðan
3sg-inan	mis	mis
3pl	luð	luð
int/rel-anim	kai	xai
int/rel-inan	bai	βai

Ferogh Languages

It was a change in grammar rather than a change in sounds that most greatly marked the initial divergence of Western Isthmus into Boésin and the Ferogh languages. Many of the genitive/accusative personal pronouns had merged with the nominatives in Western Isthmus, leaving the pronoun system highly defective. In Proto-Boésin, the loss of many gen/acc forms was simply tolerated, while in pre-Proto-Ferogh, it was resolved by the suffixation of postpositions. The nominative forms were left alone, while new accusatives were formed by suffixing -iθ to the nominatives, and new genitives were formed by suffixing -um to the old genitive/accusatives. This resulted in:

	NOM	ACC	GEN	DAT
1sg	da	daiθ	ðaum	ðad
1pl	gutʃ	gutʃiθ	ɣutʃum
2sg	tɨn	tɨniθ	θɨnum	ðatɨn
2pl	fe	feiθ	feum
3sg-anim	nɨ	nɨiθ	nɨum	ðan
3sg-inan	mis	misiθ	misum
3pl	luð	luðiθ	luðum
int/rel-anim	kai	kaiiθ	xaium
int/rel-inan	bai	baiiθ	βaium

Sound developments that led to Proto-Ferogh include the following.

• s > ʃ
• θ > s
• dʒ, ʒ > ɻ
• ɻ > ∅ / V_V, l_V, n_V
• ð, ɻ > r
• l > r / $C_V, V_C$

• Sequences of vowels are resolved according to this chart:

	-a	-ai	-au	-e	-i	-ɨ	-o	-oi	-u
a-	a(ː)	ai	au	ai	ai	ai	au	ai	au
ai-	ai	ai	ai	ai	ai	ai	ai	ai	ai
au-	au	ai	au	ai	ai	ai	au	ai	au
e-	e(ː)	ai	au	e(ː)	ai	ai	au	ai	au
i-	e(ː)	ai	au	i(ː)	i(ː)	i(ː)	ɨ(ː)	i(ː)	ɨ(ː)
ɨ-	a	ai	au	e	i	ɨ(ː)	o	oi	u
o-	o(ː)	oi	u(ː)	oi	oi	oi	o(ː)	oi	u(ː)
oi-	oi	oi	oi	oi	oi	oi	oi	oi	oi
u-	o(ː)	oi	u(ː)	ɨ(ː)	ɨ(ː)	ɨ(ː)	u(ː)	ɨ(ː)	u(ː)

(Long vowels only occurred in the dialect that became Feråjin.)

• e > ɛ
• o oi > ɔ ɔi
• n > ∅ / finally after unstressed vowels, except in certain suffixes

• Some pronouns and other grammatical forms were contracted to monosyllables, but not all.

• The remaining dative forms became used for both singular and plural, instead of being restricted to singular as in Western Isthmus.

• The dative forms also took on a locative function; and for 3sg-inanimate, 3-pl, and interrogative/relative, the genitive pronouns took on the additional functions of both dative and locative.

After these sound changes, the Proto-Ferogh pronouns were as follows:

	NOM	ACC	GEN	DAT
1sg	da	dais	raum	rad
1pl	gutʃ	gutʃis	ɣutʃum	rad
2sg	tɨn	tɨnis	sum	ratɨ
2pl	fɛ	fais	faum	ratɨ
3sg-anim	nɨ	nis	num	ran
3sg-inan	miʃ	mis	miʃum	miʃum
3pl	lur	luris	lum	lum
int/rel-anim	kai	kais	xaim	xaim
int/rel-inan	bai	bais	βaim	βaim

Faraghin

The development of Faraghin from Proto-Ferogh included the following sound changes:

• ɛ, ɔ > a
• i > e
• ɨ > i
• ai ɔi > oi
• au > eu
• β > b
• t > tʃ / _i
• i > ∅ / tʃ_#
• u > o / except before n or labials, or after the second syllable

• x > k / _s,ʃ (maybe also ɣ > g)
• t + s > tʃ / V_V (but remains ts / _#, and in loans from Ndak Ta)
• ʃ + s > s
• t > ∅ / _m,n
• sporadic syncope of unstressed vowels

Thus, the personal pronouns of Faraghin (in phonemic transcription) are:

	NOM	ACC	GEN	DAT
1sg	da	dois	reum	rad
1pl	gotʃ	gotʃes	ɣotʃom	rad
2sg	tʃin	tʃines	som	ratʃ
2pl	fa	fois	feum	ratʃ
3sg-anim	ni	nes	nom	ran
3sg-inan	meʃ	mes	meʃom	meʃom
3pl	lor	lores	lom	lom
int/rel-anim	koi	kois	xoim	xoim
int/rel-inan	boi	bois	boim	boim

Feråjin

The Feråjin began as a tribal division of the Faraghin. Later their speech diverged. The following developments left Feråjin in the position of being neither clearly a dialect of Faraghin nor clearly a separate language.

• ɔ(ː) ɔi > ɒ(ː) ɒi
• u(ː) > o(ː)
• ɨ(ː) > u(ː) (stressed)
• ɨ(ː) > i(ː) (unstressed)
• ɛ(ː) > e(ː)
• ai > eː
• au > aː
• ɒi > ɒː
• a(ː) > æ(ː)
• β > w
• ɣ > j
• tʃ > ʃ
• Vr, Vl > Vː / _C, _#
• Vm > Vː / _#

• Separate dative forms were lost. The functions of dative and locative were taken on, as in some of Faraghin's pronouns, by the genitive form.

After these changes, the table of Feråjin pronouns was as follows:

	NOM	ACC	GEN
1sg	dæ	deːs	ræː
1pl	goʃ	gos	joʃoː
2sg	tin	tins	soː
2pl	fe	feːs	fæː
3sg-anim	ni	nis	noː
3sg-inan	miʃ	mis	miʃoː
3pl	loː	loːs	loː
int/rel-anim	keː	keːs	xeː
int/rel-inan	beː	beːs	weː

Eastern Isthmus

Developments between Proto-Isthmus and Eastern Isthmus, the last common ancestor of Doroh, Kennan, Kietek, and Ka'alikora, include the following:

• p t ʈ ts k > [+asp]
• b d ɖ dz g > p t ʈ ts k / [-voice]_, _[-voice], _# (but no aspiration gained)
• l > ɬ / [+asp]_
• n l s > ɳ ɭ ʂ / [+rflex]_, _[+rflex]
• mf ms ns > mpf mps nts
• a > o / _C+(o,u) when unstressed
• C > Cʷ / _u, except /j/ was not labialized
• "V > V: / _[+vcd], unless _CC$ or _CC+ or part of a diphthong other than jV
• (s)C > (sʲ)Cʲ / _j, _i (swallowing the j) except where C = /ɬ/ or a retroflex
• jC > Cʲ / _# with the same exceptions
• i: u: > ij uj
• ij uj > ej oj

	NOM	GEN/ACC
1sg	da	asta
1pl	gʷuʈ	oskʷuʈʰ
2sg	tʰʷunʲ	ostʰʷunʲ
2pl	fe	asfe
3sg-anim	nʲo	osʲnʲo
3sg-inan	mʲis	asʲmʲis
3pl	lʷuts	oslʷuts
int/rel-anim	kʰej	askʰej
int/rel-inan	bej	aspej

Doroh

Phonetic developments leading to Proto-Doroh most notably include lenition of intervocalic plosives on the one hand and the appearance of front rounded vowels on the other. The latter appears to have been an areal development that Doroh has in common with its neighbours Lotoka and Affanonic.

• ʰ > ∅ / _#
• f > ʷ / C_
• s > h / V_C, except where the s is palatalized
• hl > ɬ
• b d ɖ dz g > β ɾ ɻ z ɣ / V_(ʲ,ʷ)V
• p t ʈ k > b d ɖ g / V_V
• n l > ɲ j / _ʲ (but not if preceded by sʲ zʲ)
• s z > ʃ ʒ / _(i,ʲ)
• ʂ sʷ > ʂʷ
• ɻ ɾʷ > ɻʷ
• zʷ > ʐʷ
• tsʷ > ʈʂʷ
• dzʷ > ɖʐʷ
• Cʲa Cʷa > Cʲe Cʷo
• Cʲo Cʷe > C(ʲ,ʷ)ø
• Cʲu Cʷi > C(ʲ,ʷ)y
• pʲ bʲ tʲ dʲ > ps bz ts dz
• tsʲ dzʲ kʲ gʲ > tʃ dʒ tʃ dʒ
• ɾʲ ɣʲ > j
• fʲ βʲ > ʃ ʒ
• pʷ bʷ > pf β
• βʷ ɣʷ > w
• ʲ ʷ > ∅
• dz dʒ ɖʐ > z ʒ ʐ / #_, C[+voice]_

The development from Eastern Isthmus to Proto-Doroh is marked by a proliferation of suffixed case forms, with the old ACC/GEN becoming the accusative, but also forming the stem of the oblique cases, except for the locative.

Following standard Dorological practice, retroflexes are written with an underdot, /ɬ/ is written ł.

	NOM	ACC	GEN	BEN	INST	COM	LOC	LAT	ABL
1sg	da	ahta	ahtajin	ahtała	ahtawum	ahtamina	dajits	ahtajila	ahtanogum
1pl	guṭ	ohkuṭ	ohkuḍin	ohkuḍła	ohkuḍum	ohkuṭmina	guḍits	ohkuḍila	ohkuṇogum
2sg	tuñ	ohtuñ	ohtuñin	ohtuñeła	ohtuñum	ohtuñemina	tuñits	ohtuñila	ohtuñogum
2pl	fe	aṣø	aṣøjin	aṣøła	aṣøwum	aṣømina	fejits	aṣøjila	aṣønogum
3sg-anim	ñø	ošnø	ošnøjin	ošnøła	ošnøwum	ošnømina	ñøjits	ošnøjila	ošnøgum
3sg-inan	mis	ašmis	ašmišin	ašmiła	ašmiṣum	ašmihmina	mišits	ašmišila	ašmihnogum
3pl	luts	ołuts	ołučin	ołutła	ołuṭṣum	ołutsmina†	lučits	ołučila	ołutsnogum†
int/rel-anim	kej	ahkej	ahkejin	ahkejła	ahkejum	ahkejmina	kejits	ahkejila	ahkejnogum
int/rel-inan	bej	ahpej	ahpejin	ahpejła	ahpejum	ahpejmina	bejits	ahpejila	ahpejnogum

†Or ołuhmina, ołuhnogum.

Central Doroh

This is a somewhat idealized representation of the pronoun forms of the central cluster of Doroh dialects, ca. -500 YP.

Preaspiration on plosives simplified to geminates by -500. This change also occurs in Western dialects, though at a later date. In the case of the Central group only, /hm hn/ simplify to voiceless /m̥ n̥/.

The animacy distinction collapsed, with the kai and bai series being reassigned according to grammatical role, and the ñø and mis series retained as competing forms with the same meaning.

The comitative case fell out of general use and was replaced by the instrumental, but it remained for singular personal pronouns only.

Finally, the 2pl began to take on a formal singular role alongside its original plural sense. The comitative aṣømina was only used in this singular role.

	NOM	ACC	GEN	BEN	INST	COM	LOC	LAT	ABL
1sg	da	atta	attažin	attała	attavum	attamina	dažits	attažila	attanūm
1pl	guṭ	okkuṭ	okkuḍin	okkuła	okkuḍum	—	guḍits	okkuḍila	okkunūm
2sg inf.	tuñ	ottuñ	ottuñin	ottuñeła	ottuñum	ottuñemina	tuñits	ottuñila	ottuñūm
2pl	ve	aṣø	aṣøžin	aṣøła	aṣøvum	(aṣømina)	vežits	aṣøžila	aṣønūm
3sg-A	ñø	øšnø	øšnøžin	øšnøła	øšnøvum	øšnømina	ñøžits	øšnøžila	øšnȳm
3sg-B	mis	ašmis	ašmišin	ašmiła	ašmišum	ašmim̥ina	mišits	ašmišila	ašmin̥ūm
3pl	luts	ołuts	ołučin	ołutła	ołuṭṣum	—	lučits	ołučila	ołun̥ūm
interrogative	kai	akkai	akkežin	akkałja	akkežum	—	kežits	akkežila	akkanjūm
relative	bai	appai	appežin	appałja	appežum	—	bežits	appežila	appanjūm

Ṭømjuñar

Ṭømjuñar was a widely-spoken Central Doroh dialect originating in the city-state of Ẓaṛmott. This describes the state of the language ca. 800.

The 3sg pronouns resolved in favor of the mih form. The original 2pl form ve supplanted the singular, while a new suffixed form venah supplied the new 2pl.

	NOM	ACC	GEN	BEN	INST	COM	LOC	LAT	ABL
1sg	da	atta	attarin	attaþ	attavum	attamin	daris	attaril	attanum
1pl	guṛ	okkuṛ	okkuṛin	okkuþ	okkuṛum	—	guṛis	okkuṛil	okkunum
2sg	ve	aṣø	aṣørin	aṣøþ	aṣøvum	aṣømin	veris	aṣøril	aṣønum
2pl	venah	aṣønah	aṣønašin	aṣønaþ	aṣønašum	—	venašis	aṣønašir	aṣønaþum
3sg	mih	ašmih	ašmišin	ašmiþ	ašmišum	ašmifin	mišis	ašmišir	ašmiþum
3pl	lus	oþus	oþučin	oþut	oþuṭum	—	lučis	oþučir	oþuþum
interrogative	kai	akkai	akkerin	akkaþi	akkerum	—	keris	akkeril	akkanjum
relative	bai	appai	apperin	appaþi	apperum	—	beris	apperil	appanjum

A Preliminary Sketch of Proto-Isthmus Grammar

This sketch summarizes most of what is currently known about Proto-Isthmus grammar. More information may be added as research progresses.

Phonology

Phoneme inventory

Consonants

	labial	dental	retroflex	palatal	velar
plosive	p b	t d	ʈ ɖ		k g
affricate		ts dz
fricative	f	s
nasal	m	n
approximant		l		j

Vowels

	front	back
close	i	u
mid	e	o
open	a

Phonotactics

Proto-Isthmus syllable structure was fairly complex, allowing up to two consonants in onsets and three in codas: (C(l,j))V((j)(l,n,s)C).

Any single consonant C could appear as onset or coda. Syllables with no onset consonant were only possible word-initially; there were no vowels in hiatus, and a single intervocalic consonant was always treated as an onset.

Onset clusters of the forms Cl Cj were found with most consonants; however, the first element C could not be a retroflex or j, and ll was not a possible onset.

Possible two-consonant coda clusters were jC lC nC sC. The final element C could not be a retroflex or j, and ll nn ss were not possible codas. Interestingly, ns sn both occured as coda clusters, and it may be that nl sl were possible too.

Three-consonant coda clusters jlC jnC jsC were also possible word-finally (with the same limitations on C as above).

In nC clusters, the n normally assimilated to the place of articulation of the following consonant (this produced allophonic [ŋ] before k g); nf may have been an exception. In sC clusters, the development in Eastern Isthmus (where voiced obstruents were devoiced after s) shows that the s must have been pronounced voiceless even when the following consonant was voiced.

When a coda and onset met, a medial cluster could form. These clusters were pretty free, but there were a few restrictions; in particular, a coda retroflex never formed a cluster with an onset approximant—if morphology brought them together, the retroflex usually mutated into a dental or velar plosive.

Stress

While stress in Isthmus has not been thoroughly studied, it seems likely that primary stress fell on the first syllable of the root or stem, leaving any prefixes unstressed.

Nouns

Proto-Isthmus nouns were inflected for number and case.

Number

Two numbers were distinguished: singular and plural. The plural was marked with a suffix which varied depending on the phonological form of the word. For the most part, nouns ending in a consonant formed their plurals by adding a suffix -o, while nouns ending in a vowel formed their plurals by changing the final vowel. The most common patterns are shown here:

stem ending	SG	PL
consonant	-C	-Co
unstressed vowel	-a, -e	-a
	-i	-ja
	-o, -u	-o
stressed vowel	-ˈV	-ˈV (no change)

Nouns already ending in -a or -o, as well as those ending in a stressed vowel, had no overt marking of the plural.

Sound changes from Proto-Eigə-Isthmus to Proto-Isthmus produced various complications. A couple of these were entirely regular: syllable constraints caused nouns ending in -Cli to form their plurals with -Cla instead of **-Clja, and (less obviously, perhaps) nouns ending in -Cji regularly formed their plurals with -Cuja. Others were synchronically unpredictable: some nouns ending in -Cju formed their plurals with -Cjo, but some used -Cja or -Cuja instead; similarly, some nouns ending in -Clu had plurals with -Clo but some used -Cla instead.

A number of nouns had consonant and/or vowel alternations in the plural. The most common of these affected nouns ending in the retroflex stops -ʈ -ɖ: before the plural suffix, the retroflexes became dentals (-to -do) in some nouns and velars (-ko -go) in others. Some nouns ending in -ʈu -ɖu had a related alternation in which the retroflexes became labials (-po -bo) in the plural. Other consonant alternations behaved similarly, but were less common (and less predictable): these included -s → -tso, -su → -fo, -n → -jno, and -nu → -mo.

Moreover, when any of these alternating consonants was preceded by the vowel i, this vowel regularly changed as well, becoming ju in the plural—unless there was already an onset cluster immediately preceding it, in which case it became u. Less predictably, in some (but not all) nouns that contained the sequence [labial onset]+[rounded back vowel]+[alternating consonant] another vowel alternation applied in the plural: u → ju and o → e. All three of these vowel alternation patterns also occured in some (otherwise regular) nouns ending in -s, -n, or -l.

Finally, some nouns ending in -j formed their plurals with -jno (rather than -jo), and some vowel-final nouns added -dzo in the plural instead of changing the final vowel. These changes were not regularly accompanied by vowel alternations.

Some of these inherited irregularities were probably smoothed away by analogy, either in Proto-Isthmus itself or in its descendants. More data and examples are needed!

(In Faraghin some plurals were lexicalized with a collective meaning: tekh 'sword' → tekha 'army' is an example, from Proto-Isthmus tisk → tisko.)

Case

Four cases were distinguished in Proto-Isthmus: nominative, accusative, genitive, and dative. Case was marked with prefixes; the paradigms fall into a small number of declension classes, defined mostly by animacy:

		NOM	ACC	GEN	DAT
Pronouns	1SG, 2SG, 3SG.ANIM	–	as-	as-	dza-
	3SG.INAN, PL, INT/REL	–	as-	as-	?
Nouns	ANIM	–	?	as-	? dz(a)-
	INAN	–	–	as-	? dz(a)-

The accusative case had no marker of its own; it was syncretic with the genitive for pronouns, and with the nominative for inanimate nouns. Animate nouns must have followed one of these two patterns, but it's probably impossible to tell which: in some descendant languages they behave like the pronouns, while in others they behave like inanimate nouns, and either pattern could easily have spread by analogy.

(Both the accusative and genitive uses of as- seem to have originated in Pre-Proto-Isthmus from an earlier dative or locative function.)

The vowel in the dative prefix dz(a)- was probably deleted before another vowel; however, this detail is uncertain since there are only a few reflexes of dative forms to compare. In fact, even the actual function of this prefix in Proto-Isthmus is uncertain; it may have been a locative or allative prefix rather than a true dative (see the discussion above).

(While the case prefixes are no longer productive in the Ferogh languages, there are a few Faraghin nouns that descend from old lexicalized genitives; in fact, Faragh is an example, from Proto-Isthmus as-pes-dosg 'of the camp-people'.)

Gender

A third category affecting nouns in Proto-Isthmus was gender: all nouns were inherently either animate or inanimate. Gender assignment was semantic—that is, which gender a noun belonged to was predictable from its meaning, with at most a handful of exceptions. Gender may not have affected the behavior of the nouns themselves (unless animate nouns formed the accusative case with as- like the pronouns), but it did affect agreement: pronouns and verbs had distinct animate and inanimate forms in the third person singular, and there may also have been adjective agreement.

Derivation

A few derivational processes affecting nouns can be traced to Proto-Isthmus.

Diminutive and Augmentative

Diminutive and augmentative nouns were formed with suffixes:

DIM	-(a)ɖu
AUG	-(a)t

The linking vowel (a) was used when these suffixes followed a consonant.

(Traces of the diminutive suffix can be seen in some Faraghin kinship terms, for example:

• Faraghin badeu 'father' < Proto-Isthmus [V]bad-aɖu 'little father'
• deu 'daughter' < [V]da-ɖu '? little girl' or '? little daughter'

The augmentative suffix can be seen in

• Faraghin kert 'anger' < Proto-Isthmus [V]kisd-at '? great violence'
• khunt 'devil' < as-kun-at '? great malevolent one'
• mašt 'house, palace' < mas-at 'great house'
• nagat 'lord, baron' < naga-t 'great lord'

and maybe in sat 'gold' < tso-t '?? great gold'.)

Verbalization

It appears that Proto-Isthmus nouns could be treated as verbs simply by adding verbal morphology to the noun stem, as one way of forming predicate nominals. There are some hints of similar behavior in Meshi and Old Eastern Miwan, so this way of forming predicate nominals may be inherited from Proto-Eigə-Isthmus.

(In Faraghin, causative verbs can be derived from nouns by adding the causative suffix -oi, followed by a regular verbal agreement suffix; there are several examples in the lexicon:

• bar 'master' → baroin 'appoint, bestow'
• ben 'nothing' → benoin 'annul'
• bran 'front, face' → branoin 'advance, march'
• tat 'law, rule' → tatoin 'legalize; permit, allow'

This suffix -oi reflects Proto-Isthmus -aj. It's not clear whether the active voice can be similarly applied to nouns in Faraghin.)

Adjectivization

Proto-Eigə-Isthmus had a suffix -ʔa or -ha that derived adjectives from nouns; but (aside from the participle suffix -dja, derived from the verbal noun suffix -di + -ʔa/-ha) it's not known whether this suffix remained productive in Proto-Isthmus. If it did, it would have the form -a in most environments.

Two later derivational suffixes that form adjectives are known from the Faraghin evidence: -(i)ts and -ujn (> Far. -s and -in). Both were probably grammaticalized after the Proto-Isthmus stage, but it's convenient to discuss them here.

• The older of these was probably -(i)ts (glossed 'quality' in the Faraghin wordlist). It derived adjectives from various parts of speech, including nouns, verbs, and other adjectives. This suffix comes from the postposition its 'on'; it is tentatively dated to the Western Isthmus stage.

• Probably somewhat younger was -ujn, which mainly derived adjectives from nouns. In the Proto-Isthmus lexicon it's treated as a separate word, tentatively glossed 'like'; perhaps it originally meant something like 'shape, likeness'.

These two suffixes are similar in many ways. Their semantics appear to be nearly identical; both are sometimes glossed '-ish' in the Faraghin wordlist. Also, both of them sometimes seem to work in reverse in Faraghin, deriving proper nouns from adjectives:

• čints 'on the left, north' → Čintsin 'the northern mountains'
• muns 'soft, weak' → Munsin 'weakish; the Ngauro'
• porat 'clean, pure' → Porats 'the river dividing Faraghin and Feråjin territory'

In fact, the use of unmarked adjectives as nouns is perfectly regular (see below); in each of these examples the suffix seems to function as a semantically empty derivation that, without directly changing the adjective's meaning, produces a modified form of the word that can then be used in a narrowed sense as a proper noun.

Compounding

Two types of compound nouns are attested:

• Adjective-Noun: Faraghin khoir 'goat' < Proto-Isthmus as-koɖ[V] 'good' + idz 'head; head of livestock, domesticated animal'

• Noun-Genitive: Faraghin nagho 'bear' < Proto-Isthmus nag(a) 'lord' + su[n] '(of the) forest'

Exactly the same compound patterns are also found in Forest Miwan, suggesting that they may be inherited from Proto-Eigǝ-Isthmus.

Pronouns

The personal pronouns of Proto-Isthmus were as follows:

	NOM	GEN/ACC	DAT
1SG	da	asda	dzada
1PL	guʈ	asguʈ	?
2SG	tujn	astujn	dzatujn
2PL	fe	asfe	?
3SG.ANIM	njo	asnjo	dzanjo
3SG.INAN	mis	asmis	?
3PL	ludz	asludz	?
INT/REL.ANIM	kej	askej	?
INT/REL.INAN	bej	asbej	?

Though we can posit the forms of the missing dative pronouns (e.g. **dzaguʈ), no reflexes of such forms can be identified in any Isthmus language (see the discussion above).

Verbs

Proto-Isthmus verbs were inflected for aspect, voice, and subject agreement. Verbal inflection is both more complex and less well-understood than nominal morphology, and many details of what follows are uncertain.

The structure of the Proto-Isthmus verb can be described by this template:

(prefix or reduplication)-

INNER STEM

-aspect

-(stem consonant)

-voice

-subject agreement

Here are a couple of examples of fully inflected verbs:

asgospumis

as-go<s>p-u-mis

PREFIX-take<RES>-DETR-3SG.INAN

‘it’s been stolen’

dadajaludz

da~daj-a-ludz

ITER~dance-ACT-3PL

‘they danced and danced’

Verbal Prefix(es)

Proto-Isthmus had at least one verbal prefix of the form Vs-. While this has been identified with the adjective-intensifying prefix as- (see Intensification below), with verbs the prefix's function seems less clear: cf. as-gosp- 'steal', as-kjustil- 'bequeath', as-pusg- 'make, craft', and maybe as-gont- 'fly'. Perhaps there were multiple Vs- prefixes that we haven't quite sorted out yet.

Aspect

Aspect marking in Proto-Isthmus is probably best thought of as derivational (with the possible exception of the imperfective/durative n-grade, which is said to form a consistent aspectual operation in Kennan). The values marked included perfective/punctual, imperfective/durative, resultative, inceptive/inchoative, and iterative/intensive. (These names are intentionally slightly vague; the precise meanings that most of these categories had at the Proto-Isthmus stage are not known for certain.)

Forms: the Grades

The forms that expressed these categories are traditionally called grades, and are a somewhat mixed bag. The iterative/intensive was formed by reduplicating the first CV of the verb, while most of the others were formed with infixes in some verbs and suffixes in others—thus forming two major inflectional classes, infixing verbs and suffixing verbs.

The forms found in Proto-Isthmus included the following:

	Grade	Infixing	Suffixing
Perfective/Punctual	Zero-grade	-	-
Imperfective/Durative	N-grade	<(i)n>	-(i)n
Resultative	S-grade	<(u)s>	-(u)s
Inceptive/Inchoative	J-grade	<i>, <j>	-j
Iterative/Intensive	Reduplicated grade	REDUP~	REDUP~

There may have been other grades as well; the possibilities include an l-grade (similar to the n- and s-grades) and maybe a retroflex grade (which would turn root-final stops into retroflexes). It's not known what categories such grades might have marked.

Various inflectional subclasses probably existed. At the least, both infixes and suffixes had syllabic and non-syllabic allomorphs (though the j-grade suffix was probably never syllabic). The syllabic forms of the n-grade and s-grade might each have had several variants.

Functions

As noted above, the n-grade may have been inflectional in Proto-Isthmus; if so, it probably marked a basic imperfective aspect, contrasting with the perfective zero-grade. (There is one possible example of a noun derived from an n-grade: Faraghin ghantač 'dragon', apparently cognate with Miwan kwintas 'bird'; but at present it's not clear whether the n in this word is an infix or part of the root.)

The s-grade could be used to derive resultative nouns, a process which was highly productive in Proto-Isthmus. S-grade forms could also be used as verbs: examples of this include dog- 'set down' → dosg- 'camp, stay', gop- 'take, carry' → (as)-gosp- 'steal', pleg- 'meet' → plesg- 'get to know, get or be used to', and perhaps meg- 'boil, behave wildly or badly' → mesg- 'raid, capture'.

The j-grade derived inceptive or inchoative verbs, i.e. verbs denoting the beginning of an action or (especially) a state. Examples include [V]dludz- 'be awake' → [V]dludzj- 'awaken', dzusn- 'be red' → dzusin- 'turn red', plun- (perhaps meaning 'cut' or 'cut up') → plunj- 'stab, slash', and possibly tlub- 'be married' → tlujb- 'get married'.

The reduplicated grade formed iterative or intensive verbs; the sole example currently in the lexicon is daj- 'dance' → da~daj- 'dance and dance, dance a lot or intensely'.

Voice

Voice marking in Proto-Isthmus is not completely understood yet, but the system was probably simpler than in its ancestor Proto-Eigə-Isthmus. The primary voices were active, causative, and detransitive; in PEI these could be combined with one another to form compound voices such as detransitive of causative and causative of detransitive. The full set of inherited voice suffixes (traditionally called stem vowels) would have looked like this:

Active	-a
Causative	-i, -aj
Detransitive	-u
Detransitive of Causative	-ju
Causative of Detransitive	-uj

But not all of these forms remained productive, perhaps even in Proto-Isthmus. In the branch of Isthmus that led to Faraghin, the simple causative suffix -i was replaced by the compound suffix -aj (except for a few fossilized forms such as dim-i- 'give'), while the detransitive suffix -u seems to be preserved only in the participle. It's not yet known at what stage the basic causative and detransitive forms were abandoned. Moreover, the detransitive of causative and causative of detransitive forms are entirely unattested in Faraghin, and it seems likely that they were not in fact found in Proto-Isthmus at all.

The exact function of the detransitive voice in Proto-Isthmus is unknown; it could have been a passive, an anticausative, a middle voice, or a general intransitivizer with several functions. (The Faraghin reflex appears to be a passive participle, but passives can easily develop from other kinds of intransitivizers.)

Subject Agreement

Verbs agreed with their subjects in person, number, and probably (for third person singulars) gender. The first and second person forms have not been determined yet; but the third person forms were probably marked with suffixes that were identical to the pronouns:

3SG.ANIM	-njo
3SG.INAN	-mis
3PL	-ludz

Deverbalization

The Participle

Proto-Isthmus had a verbal adjective, the participle, formed with the suffix -dja. The participle was inflected for voice; it may be helpful to list the voice forms separately:

Active	-a-dja
Causative	-i-dja, -aj-dja
Detransitive	-u-dja

Here are some examples:

Active:

ɖujfadja

ɖujf-a-dja

leap-ACT-PTCP

askoɖidz

as-koɖ-idz

INTENS-good-domesticated.animal

‘leaping goat’

lupadja

lup-a-dja

deceive-ACT-PTCP

askunat

askun-at

devil-AUG

‘lying devil’

plunjadja

plun-j-a-dja

cut.up-INCHO-ACT-PTCP

tisko

tisk-o

sword-PL

‘slashing swords’

Causative:

dadajidja

da~daj-i-dja

ITER-dance-CAUS-PTCP

Bleɖus

Bleɖus

Fate

‘Fate who makes (us) dance’

nopolidja

nopol-i-dja

be.bored-CAUS-PTCP

taʈa

taʈa

sister

astujn

as-tujn

GEN-2SG

‘your boring sister’

kijbajdja

kijb-aj-dja

eat-CAUS-PTCP

badaɖu

bad-aɖu

father-DIM

asludz

as-ludz

GEN-3PL

‘their little father who feeds (them)’

Detransitive:

asgospudja

as-go<s>p-u-dja

PREFIX-take<RES>-DETR-PTCP

tsot

tso-t

gold-AUG

‘stolen gold’

lupudja

lup-u-dja

deceive-DETR-PTCP

nagat

naga-t

lord-AUG

‘deceived lord’

dzusinudja

dzus<i>n-u-dja

red<INCHO>-DETR-PTCP

asgujtsak

as-gujts-ak

INTENS-sharp-NMLZ

‘reddened knife’

Like other adjectives, the participle could be used as a noun with zero-derivation (see De-adjectivization below); this could be used to form agentive and patientive nouns. (The detransitive participle is the only form currently attested in Faraghin, whose lexicon includes several lexicalized examples.)

The Verbal Noun

There was probably also a verbal noun formed with the related suffix -di; this noun may have been inflected for voice much like the participle. (The cognate form in Ngauro and Old Eastern Miwan is an action noun formed with -ti; this was borrowed in early Faraghin as a nominalizing suffix -(a)č.)

The Resultative Noun

An important use of the s-grade was to derive a resultative noun, a nominalizing process which was highly productive in Proto-Isthmus. This noun was not inflected for voice; examples include gad- 'flow' → gasd 'stream', kid- 'be angry' → kisd 'anger', sud- 'go around, move in a circle' → susd 'circle, year', tsik- 'to call, to name' → tsisk 'name', and ʈuk- 'squeeze, clench' → ʈusk 'fist'.

Adjectives

Little is known about Proto-Isthmus adjectives so far. They seem to have behaved like verbs in some ways (taking verbal morphology in some predicative constructions, for example), while in other ways they seem somewhat noun-like.

Derivation

We do know a little about derivational processes involving adjectives.

Intensification

Proto-Isthmus had an adjective-intensifying prefix as-. Some intensified adjectives came to replace the corresponding unmarked adjective: e.g. Faraghin khar 'good' from Proto-Isthmus as-koɖ[V] 'very good'.

Nominalization

An attributive adjective could stand as the head of a noun phrase when the modified noun was understood or unspecified; in the Faraghin wordlist several examples are found where this has become lexicalized as a noun (e.g. českod 'hidden; hidden one; mythical pixie-like being'). There's a particularly notable instance where a Faraghin participle with a noun-like definition takes the regular negative prefix used for adjectives:

• rantan 'test (by experience), prove' → rantod 'veteran, experienced man' → fisrantud 'untried person, rookie'

(Cf. tats 'lawful, legal' → fistats 'unlawful, illegal'.) So rantod seems to carry its adjectival morphology with it, even when it's used in this noun-like way.

Similar behavior can be observed in Ngauro (e.g. dyaka 'kingly, royal; royal one, (his/her) majesty', borrowed into Ndak Ta as diàka 'king, ruler'), suggesting that this may be a shared inheritence from Proto-Eigə-Isthmus.

There's also an old and rather obscure suffix -(a)k that seems to derive nouns from adjectives. An example of this is Faraghin ghisk 'knife', apparently from Proto-Isthmus gujts 'sharp' + -ak; Faraghin českan 'to hide' may be another, if it comes from an adjective ʈits (perhaps meaning 'over, covering') + -ak + verbal suffixes. This suffix seems to have become unproductive quite early in the prehistory of Faraghin (perhaps even by the Western Isthmus stage).

Verbalization

It seems that Proto-isthmus adjectives, like nouns, could be treated as verbs simply by adding verbal morphology, as one way of forming predicate adjectives.

(In the Faraghin lexicon, adjectives are found with regular causative verb morphology:

• leus 'free' → leusoin 'liberate'
• naks 'outside (adj.)' → naksoin 'exile, cast out'
• šes 'holy' → šesoin 'anoint; ordain; crown'

There's also rušnen 'paint, color; insult, accuse' < PIsth. dzus<i>n-i- 'redden (something); paint, color', which has an older form of causative marking that's no longer productive in Faraghin. It must date to before Western Isthmus, and it's cognate with Eastern Miwan tun 'red'. This shows both that this pattern of causative verbs formed directly from adjectives goes back to Proto-Isthmus, and that the grades/infixes could also be applied to adjectives in this way.)

Syntax

Very little is known about Proto-Isthmus syntax at present; this section briefly notes a few bits that can be inferred from the current evidence.

Word Order

It appears that Proto-Isthmus word order in general tended to be head-final: Proto-Isthmus had postpositions, adjectives preceded the nouns they modified, and auxiliaries apparently followed their main verbs (at least in the variety that led to Faraghin). However, its ancestor Proto-Eigə-Isthmus seems to have had a number of head-initial structures, and traces of these survived into Proto-Isthmus: for example, the case prefixes and noun-genitive compounds mentioned above. It's also possible that Proto-Isthmus kept some prepositions alongside its postpositions, or allowed genitives to follow their possessed nouns in full noun phrases (that is, not just in compounds).

Postverbal auxiliaries are typical of OV languages, and on this basis it's been posited that Proto-Isthmus had a basic SOV order. However, if these auxiliary constructions developed later (perhaps in Western Isthmus or Proto-Ferogh) the Proto-Isthmus order could have been different.

Postpositions

A few Isthmus postpositions are known so far (because they became case suffixes in daughter languages).

• um 'of' became a genitive suffix in Ferogh. This morpheme dates back to Proto-Eigə-Isthmus, where it was probably a preposition (its cognate in the Eigǝ Valley languages is a prefix m-); how its use differed from that of the genitive prefix as- is unknown.

Some of the other postpositions may have been grammaticalized later from nouns and/or verbs; it's possible that a few of them didn't develop until after the Proto-Isthmus stage.

• its 'on' (probably related to idz 'head') became both an adjectivizing suffix (see Adjectivization above) and an accusative suffix in Ferogh, and a locative suffix in Doroh. Its locational meaning can be seen in a few older derivations, e.g. tjunt 'left hand' → tjunt-its 'on the left, north', and probably gasd 'stream' → gasd-its 'boat'; Faraghin lets 'longsword' may be similarly derived from lit (a noun of unknown meaning) + its.
• in (original meaning unknown) became a genitive suffix in Doroh.
• mina 'with, together with' became a comitative suffix in Doroh.
• ila 'to, toward' became an allative suffix in Doroh.
• asla 'for, for the sake of' (possibly related to ila) became a benefactive suffix in Doroh.
• nak 'outside' became an extraessive suffix in Faraghin.

Auxiliary Verb Constructions

Periphrastic constructions may have been used to express additional verbal distinctions (such as tense, aspect, and/or modal categories); certainly such constructions existed in later Isthmus languages. One example is represented by the Faraghin iterative suffix -čan, which appears to be a contracted form of an earlier auxiliary verb. At present it's not known whether this verb was used as an auxiliary as early as Proto-Isthmus, but it certainly could have been.

Negation

Proto-Isthmus adjectives were negated by adding a morpheme pjuts[V] or (with what appears to be the intensifying prefix) aspjuts[V] immediately before the adjective. This morpheme became a prefix (fis-) in Faraghin, but may have been a separate word in Proto-Isthmus.

It's not known whether this construction was also used to negate clauses.

Proto-Isthmus Lexicon

See Proto-Isthmus/Lexicon