Proto-Samic

Proto-Samic is the reconstructed last common ancestor of the Samic languages. It was spoken in the 1st millennium CE in the area of modern-day northern Finland. (If extinct southern languages are included, this dating will likely become both older and more southern. Stem-group para-Samic languages have from toponymic evidence once been spoken as far south as the Baltic coast)

Sound changes from Proto-Uralic

Possibly original:

• e → ɪ / _Ci (see *ê)
• e → e / _{j w x} (ie. does not become *ɪ → *ɤ)
• Vx → Və → Vː / _C

Proto-Uralic to Pre-Samic (≈ "Finno-Volgaic")

Shared with Finnic & Mordvinic

• ë → a
• ðʲ → ð (also lʲ → l, if it exists separately; note that other palatals remain)
  • NB: Sammallahti posits ðʲ → θ / V_C (*käðʲwa "weasel" → *káθwé → NS gádfe - but can this be ruled out for ð? Cf. *kotwa > IS kuáđfi)

Shared with Mordvinic

• ü → i (feeds the next changes)
• i → iː / _{j x}
• u → uː / _{w x}
• i u → ɪ ʊ (*iː *uː remain) (also shared with Ugric, Nenets!)

Pre-Samic to Proto-Samic

• V → Vː / [-HIGH] (also shared with Nenets!)

a-umlaut

• eː äː oː → ea aː oa / _C{aː, oː} (note that this *aː is distinct from *ɑː!)

• x, ŋ → w / _t (*xt by now only occurring in newer derivativs/loans)
• ? ŋ → w / _ć (*(k/s)lāvćā ~ H. légy, or ONo kleggi)

Spirant fortition

• ś x → ć k / _V
• PS *ć (written also *č) is perhaps describable as a palatal stop, so we may be dealing with a process [sʲ ~ ɕ] → [ç] → [c]. Note that original *s *š remain sibilants!
• PS *ś (likewise also written *š) remains elsewhere
  • śj, sj → śś
  • ðj → ðś
• ð → θ / #_ (South/Ume h-, elsewhere merges with *t-)

Fronting of postalveolars

• tʃ ʃ <č š> → ts s <c s>

'Preaspiration

• Pː → ʰPː (both grades; affricates as well)

Popular areal changes

• w → ∅ / _{y u o} (shared with Finnic, Mordvinic, Mari)
• j → ∅ / _{i e} (not sure about the distribution of this)
• w → v (even syllable-finally)

Raising & epenthesis (first two shared with North Finnic but obviously non-genetic)

• äː ɑː → eː oː / [+STR]
• ∅ → v / #_oː
• ∅ → j / #_eː
• aː → eː / [-STR]
       except remains / [-STR]_ɪ
        → iː / [-STR]_j
• oː → uː / [-STR]_ɪ

Gradation could be described around here?

Cluster stuff

• tr → rtt (in loans)
• k → v / _j (*rōvjē *vōvjē, possibly *tāvjē if not from Fi. *taɣja; South S. has -kś-)
       _l (*keavlē *kōvlē- *nāvlē *tōvlē; *ćīvlē with *ük > *iv > *īv ?!)
       _ń (*ćóvné)
• t → r / _n (*vōrnē-; but PGmc *bodam → *ponné, later loan?)
• n → s / _s (*kāssē *koassē *kōssē *peassē *vōssē)
• ŋ → k / _s (*jōksɤ)

Centralization (a late change - consider eg the northern Lappland toponym *ɤnārɤš → Finnish Inari)

• ɪ ʊ → ɤ o

There may have also been a high lax unrounded back vowel (see ô_a). A merger into that might provide the motivation for the unusual back reflex of PU *i.

---

This results in the following inventory:

• long vowels */iː eː ea aː oa oː uː/ <ī ē ea ā oa ō ū>
• short vowels */ɤ o/ <ë o>

/ea oa/ are traditionally analyzed as [ɛː ɔː] rather than [ea oa], however I'm at a loss to explain how would *ɑː end up as *oː in that case.

The shortening of iː uː → i u, the backing of e → ɤ and the diphthongization eː oː → ie uo (when stress'd) may postdate Proto-Samic; the former two have bimoraic/front reflexes /ij uv e/ in South Samic. At any rate, as eg. *a *e in Scandinavian loans ends up as *ā *ē, the "long" vowels must've been the unmarked series.

Sound changes to individual Sami languages

Proto-Samic to Proto-Western Samic

• preaspiration spreds to strong grade simple stops/affricates
• ŋm → mː
• ŋv → vk ([vg̊]) (versus East *vː, so why not just PS *vː?)
• NP → Bː (also affricates; later also in Inari, Skolt, Kemi, possibly South; cf. Scandinavian)
• ś → jʰ / _{n t k} (counterfeeds medial voicing in South)
• Nː → ʔN (orthographic <bm dn>; counterfeeds gemination in South)

PWS to South Samic

• θ ð → h r
• Short vowel split
  • ë o ī ū → ɨC uC ɨj uv / open 1st syllable (+ gemination after *ë *o)
  • ë o ī ū → a a ɨ u / closed 1st syllable
  • Possibly works better if preceding the lowering *i *u → *ë *o (!!)
• ē ea → ië ie
• ē ë ō → ie e u [-STR] (& more complications, but this'll do)
• umlauts
  • ɨ u a → i y ä / _C*ie
  • a → ä / _C*u (if from *ë)
  • ɨ a → y o / _v
• medial voicing of plain stops

Proto-W Samic to Proto-NW Samic

Comprises Northern, Lule & Pite Sami.

• ɤ → ɑ (also Inari; some dialects historically had ɤ → e / _C*{i u})
• m → b̥ / _P[-LAB] (*b̥d̥ *b̥ć̥ remain distinct from *pt *pć; also *ćɤmkɤ → *ćɑpkɑ; partially also in South)
    nbsp; nbsp; →
• N₁ → P / _N₂ (ie. nm mn → tn pm)
• "kŋ" → ʔŋ (*tōkŋɤ, *ākŋō; South ŋ, Inari/Kildin/ vŋ, Skolt ŋː. Mari/Komi #tum-?)
• ck ćk ćm → sk śk śm (also Ume, Inari, Kemi, northern Skolt; check other affricate clusters?)
• strong grade of singletons merges with weak grade of geminates (also Inari, Skolt, Akkala; mostly also Ume, but four degrees attested dialectally!!)
• repeat: Nː → ʔN (later in Ume where 4 lengths attested; perhaps not common West Samic at all)
• N → ʔN / r_
• p → β / _ʰk (strong grade)
• θ → t / #_

PNWS to Lule-Pite

• vj → jv (sporadically in North, Ume, South; or does this go before *kj > vj?)
• ð → d̥ / C_

PNWS to North Sami

• ie ea oa uo → iː eː oː uː / _C*V[+hi], further → i e o u / _hC etc.
• eː oː → i u / [-STR]

(other vowel stuff probs too?)

• k → ɣ / _{s ʰt ʰć} (strong grade; also East, Lule)
• p b̥ → k / _{t c ć s ś} (both grades)
  • further → ɣ in most dialects
• ɣ → v / _C
  • Sammallahti claims v → k / _S, strong grade, but what's wrong with retention + analogy on *vs *vś?
• p [b̥] → β / _{r l ð}
  • further βr βð → rv ðβ
  • also Inari, partially Skolt, Lule
• N → T / Vs_, with sń → st
  • spred irregularly to other idioms
  • sm generally retain'd
• t [d̥] → ð / V_V (weak grade; Finnish influence? also Inari, Skolt)
• lm rd̥ jg̊ vs → lmː rd̥ː jg̊ː vsː (wtf?)
• θv → ðf (wtf? could be handled better, I'm sure)

Medial consonant clusters

Data extracted from Juhani Lehtisalo: Yhteissaamelainen sanasto (SUST vol. 200)

(proper colors/numbers TBA)

Two-consonant clusters

2nd → 1st ↓		p	t	c	ć	k	s	ś	ð	l	r	v	j	m	n	ń	ŋ	∑		Notes		Frequency color code
N	m	mp	mt	mc	mć				N/A			N/A		mm	mn			7				single root
	n		nt	nc	nć										nn			16				2-3 roots
	ŋ					ŋk	ŋs							(ŋm)				9				4-6 roots
P	p	pp	pt	pc	pć	pk	ps		pð	pl	pr			N/A				7				7-10 roots
	t		tt			tk	N/A		N/A			tv						3				11-15 roots
	c			cc		ck												3				16-20 roots
	ć				ćć	ćk								ćm
	k		kt	kc	kć	kk	ks	kś		kl		kv					(kŋ)	12
Sbl	s		st			sk	ss		N/A			sv		sm	sn		sŋ	2
	ś		śt			śk		śś						śm	śn			5
Spr	ð					ðk			ðð			ðv					ðŋ	2
L	l	lp	lt		lć	lk	ls	lś		ll		lv	lj	lm	ln	lń	lŋ	25
	r	rp	rt		rć	rk	rs	rš			rr	rv	rj	rm	rn	rń	rŋ	10
sV	v		vt	vc	vć	vk	vs	vś	vð	vl	vr	vv	vj		vn	vń	vŋ	6
	j	jp	jt			jk	js		jð	jl	jr	jv	jj	jm	jn		jŋ	6
∑		4	19	7	37	7	10	(1)	3	1	0	6	5	7	1	4

Geminate clusters

2nd → 1st ↓	pp	tt	cc	ćć	kk	∑
m	mpp					1
l	lpp	ltt		lćć	lkk	16
r		rtt (9)	rcc	rćć	rkk	18
v					vkk	7
j		jtt	jcc		jkk	10
∑	3	18	3	5	23

Comments

• *-tr- in IE loans is substituted by metathesized *-rtt-.
• *ðð is found in two words and may have been phonetically [-θð-].
• I would reconstruct *kv rather than *ŋv for the set of words showing WS *-vk-, ES *-vv-. This allows setting up a constraint against nasal+semivowel clusters & is supported by Scandinavian loan etymologies.
• Some unsecure clusters (in parentheses) may result from variable adoption of substrate loanwords:
  • "*ŋm" is based on a single word showing WS *-ŋ-, Inari -ŋ-, -vŋ- in rest of ES.
  • "*kŋ" is based on two words showing NWS *-ŋ-, versus Peninsular ES *-vŋ-. Inari and Skolt Sami have -ŋ- for the first, -vŋ- for the 2nd.