Old English Phonology - Vowel Changes in Accented Syllables

Vowel Changes in Accented Syllables

See also: Phonological history of English#Through Middle English

NOTE: Another version of this table is available at Phonological history of English#Through Middle English. This covers the same changes from a more diachronic perspective. It includes less information on the specific differences between the Anglian and West Saxon dialects of Old English, but includes much more information on the Proto-Indo-European changes leading up to the vowels below, and the Middle English vowels that resulted from them.

NOTE: This table only describes the changes in accented syllables. Vowel changes in unaccented syllables were very different and much more extensive. In general, long vowels were reduced to short vowels (and sometimes deleted entirely) and short vowels were very often deleted. All remaining vowels were reduced to only the vowels /u/, /a/ and /e/, and sometimes /o/. (/o/ also sometimes appears as a variant of unstressed /u/.)

West Germanic	Condition	Process	Old English		Examples
				i-umlaut
*a		Anglo-Frisian brightening	æ	e	*daga(z) > dæġ "day"; *batizôN > betera "better"; *taljanaN > tellan "to tell"
	+n,m		a,o	e	*mann(z), manni(z) > man, mon "man", plur. men "men"
	+nf,nþ,ns	Ingvaeonic nasal spirant law	ō	ē	*tanþ(z), tanþi(z) > tōþ, plur. tēþ "tooth"; *gans, gansi(z) > gōs "goose", plur. gēs "geese"
	(West Saxon) +h,rC,lC	breaking	ea	ie	*alda(z), aldizôN > eald "old", ieldra "older" (cf. "elder")
	(Anglian) +h	breaking, Anglian smoothing	æ	e
	(Anglian) +lC	retraction	a	æ	*alda(z), aldizôN > ald "old", ældra "older" (cf. "elder")
	(Anglian) +rc,rg,rh	breaking, Anglian smoothing	e	e
	(Anglian) +rC (C not c,g,h)	breaking	ea	e
	(West Saxon) +hV,hr,hl	breaking, h-loss	ēa	īe	*slahanaN, -iþ > slēan "to slay"; *stahlijaN > stīele "steel"
	(Anglian) +hV,hr,hl	breaking, Anglian smoothing, h-loss	ēa	ē	*slahanaN, -iþ > slēan "to slay, 3rd sing. pres. indic. slēþ "slays"; *stahlijaN > stēle "steel"
	(West Saxon) k,g,j+	palatal diphthongization	ea	ie	Lat. castra > ċeaster "town, fortress" (cf. names in "-caster, -chester"); *gasti(z) > ġiest "guest"
	before a,o,u1	a-restoration	a	(by analogy) æ	plur. *dagôs > dagas "days"; *talō > talu "tale"; *bakan, -iþ > bacanaN "to bake", 3rd sing. pres. indic. bæcþ "bakes"
	(mostly non-West-Saxon) before later a,o,u	back mutation	ea	eo2	*alu > ealu "ale"; *awī > eowu "ewe", *asilu(z) > non-West-Saxon eosol "donkey"
	before hs,ht,hþ + final -i(z)	palatal umlaut	N/A	i (occ. ie)	*nahti(z) > nieht > niht "night"
*e3			e	N/A3	*etanaN > etan "to eat"
	+m		i	N/A	*nemanaN > niman "to take"
	(West Saxon) +h,rC,lc,lh,wV	breaking	eo	N/A	*fehtanaN > feohtan "to fight"; *berkanaN > beorcan "to bark"; *werþanaN > weorðan "to become"
	(Anglian) +h,rc,rg,rh	breaking, Anglian smoothing	e	N/A	*fehtanaN > fehtan "to fight"; *berkanaN > bercan "to bark"
	(Anglian) +rC (C not c,g,h); lc,lh,wV	breaking	eo	N/A	*werþanaN > weorðan "to become"
	+hV,hr,hl	breaking, (Anglian smoothing,) h-loss	ēo	N/A	*seh(w)anaN > sēon "to see"
	+ late final hs,ht,hþ	palatal umlaut	i (occ. ie)	N/A	*sehs > siex "six"; *rehta(z) > riht "right"
	(West Saxon) k,g,j+	palatal diphthongization	ie	N/A	*skeranaN > sċieran "shear"
*i			i	i	*fiskaN > fisċ "fish"; *itiþ > 3rd sing. pres. indic. iteþ "eats"; *nimiþ > 3rd sing. pres. indic. nimeþ "takes"; *skiriþ > 3rd sing. pres. indic. sċirþ "shears"
	+ nf,nþ,ns	Ingvaeonic nasal spirant law	ī	ī	*finf > fīf "five"
	(West Saxon) +h,rC	breaking	io > eo	ie	*Pihtôs > Piohtas, Peohtas "Picts"; *lirnōjanaN > liornian, leornian "to learn"; *hirdija(z)2 > hierde "shepherd"; *wirþiþ > 3rd sing. pres. indic. wierþ "becomes"
	(Anglian) +h,rc,rg,rh	breaking, Anglian smoothing	i	i	*stihtōjanaN > stihtian "to establish"
	(Anglian) +rC (C not c,g,h)	breaking	io > eo	i	*a + firrijanaN > afirran "to remove" (cf. feorr "far")
	(West Saxon) +hV,hr,hl	breaking, h-loss	īo > ēo	īe	*twihōjanaN > twīoġan, twēon "to doubt"
	(Anglian) +hV,hr,hl	breaking, Anglian smoothing, h-loss	īo > ēo	ī	*twihōjanaN > twīoġan, twēon "to doubt"; *sih(w)iþ > 3rd sing. pres. indic. sīþ "sees"
	before w	breaking	io > eo	i	*niwulaz > *niowul, neowul "prostrate"; *spiwiz > *spiwe "vomiting"
	before a,o,u	back mutation	i (io, eo)	N/A	*miluk(z) > mioluc,meolc "milk"
*u			u	y	*sunu(z) > sunu "son"; *kuman, -iþ > cumanaN "to come", 3rd sing. pres. indic. cymþ "comes"; *guldijanaN > gyldan "to gild"
	+ nf,nþ,ns	Ingvaeonic nasal spirant law	ū	ȳ	*munþ(z) > mūþ "mouth"; *wunskijanaN > wȳsċan "wish"
	before non-nasal + a,e,o4	a-mutation	o	(by analogy) e	*guldaN > gold "gold"; *duhter, duhtri(z) > dohter "daughter", plur. dehter "daughters"
	+hV,hr,hl	h-loss	ū	ȳ	*uhumista(z) > ȳmest "highest"
(*ē >) *ā		Anglo-Frisian brightening	(West Saxon) ǣ	ǣ	*slāpanaN > slǣpan "to sleep", Lat. strāta > strǣt "street"; *dādi(z) > dǣd "deed"
			(Anglian) ē	ē	*slāpanaN > slēpan "to sleep", Lat. strāta > strēt "street"; *dādi(z) > dēd "deed"; Lat. cāseus > ċēse "cheese"; *nāha(z), nāhista(z) > nēh "near" (cf. "nigh"), superl. nēhst "nearest" (cf. "next")
	(West Saxon) k,g,j+	palatal diphthongization	ēa	īe	*jārō > ġēar "year"; Lat. cāseus > ċīese "cheese"
	+n,m		ō	ē	*mānôN > mōna "moon"; *kwāni(z) > kwēn "queen"
	(West Saxon) +h	breaking	ēa	īe	*nāha(z), nāhista(z) > nēah "near" (cf. "nigh"), superl. nīehst "nearest" (cf. "next")
	+w;ga,go,gu;la,lo,lu	a-restoration	ā	ǣ	*knāwan, -iþ > cnāwanaN "to know", 3rd sing. pres. indic. cnǣwþ "knows"
*ē			ē	ē	*mēdaN > mēd "reward"
*ō			ō	ē	*fōt(z), fōti(z) > fōt "foot", plur. fēt "feet"
*ī			ī	ī	*wībaN > wīf "wife"; *līhiþ > Anglian 3rd sing. pres. indic. līþ "lends"
	(West Saxon) +h	breaking	īo > ēo	īe	*līhanaN, -iþ > lēon "to lend", 3rd sing. pres. indic. līehþ "lends"
*ū			ū	ȳ	*mūs, mūsi(z) > mūs "mouse", plur. mȳs "mice"
*ai			ā	ǣ	*staina(z) > stān "stone", Lat. Caesar > cāsere "emperor", *hwaitijaN > hwǣte "wheat"
*au			ēa	(West Saxon) īe	*auzōN > ēare "ear"; *hauzijanaN > hīeran "to hear"; *hauh, hauhist > hēah "high", superl. hīehst "highest"
				(Anglian) ē	*auzōN > ēare "ear"; *hauzijanaN > hēran "to hear"
	(Anglian) +c,g,h;rc,rg,rh;lc,lg,lh	Anglian smoothing	ē	ē	*hauh, hauhist > hēh "high", superl. hēhst "highest"
*eu5			ēo	N/A5	*deupa(z) > dēop "deep"; *fleugōN > flēoge "fly"; *beodanaN > bēodan "to command"
	(Anglian) +c,g,h;rc,rg,rh;lc,lg,lh	Anglian smoothing	ē	N/A	*fleugōN > flēge "fly"
*iu5			N/A	(West Saxon) īe	*biudiþ > 3rd sing. pres. indic. bīett "commands"; *liuhtijanaN > līehtan "to lighten"
				(Anglian) īo	*biudiþ > 3rd sing. pres. indic. bīott "commands"
	(Anglian) +c,g,h;rc,rg,rh;lc,lg,lh	Anglian smoothing	N/A	ī	*liuhtijanaN > līhtan "to lighten"

1 The process of a-restoration, as described here, reversed the previous process of Anglo-Frisian brightening, leaving an /a/. However, it was blocked when an /i/ or /j/ followed in the next syllable; instead, /a/ was converted to /æ/ by Anglo-Frisian brightening, and then umlauted to /e/. This accounts for the outcomes of PG *talō > talu "tale" vs. the related PG *taljanaN > tellan "to tell". However, in some instances when a-restoration was blocked, the /æ/ that remained from Anglo-Frisian brightening was still reverted to /a/ by analogy with related words where a-restoration did apply; this /a/ was then umlauted to /æ/. This happened especially in verbs when some forms (e.g. the third-person singular present indicative) had umlaut, and other forms (e.g. the infinitive) did not; for example, PG *bakanaN > OE bacan "to bake" vs. PG *bakiþi > OE bæcþ "(he) bakes". This accounts for the "(by analogy)" notation in the i-umlaut column. The following diagrams show the processes involved in more detail:

No analogy:

Step	"tale"	"to tell"	Reason
1	/talō/	/taljanaN/	original forms
2	/talu/	/talljan/	after various changes, irrelevant here (e.g. West Germanic gemination)
3	/tælu/	/tælljan/	Anglo-Frisian brightening
4	/talu/	/tælljan/	a-restoration
5	/talu/	/tælljan/	unaffected by analogy
6	/talu/	/telljan/	i-mutation
7	talu	tellan	after further changes, irrelevant here

Analogy:

Step	"to bake"	"(he) bakes"	Reason
1	/bakanaN/	/bakiþi/	original forms
2	/bakan/	/bakiþ/	after various changes, irrelevant here
3	/bækan/	/bækiþ/	Anglo-Frisian brightening
4	/bakan/	/bækiþ/	a-restoration
5	/bakan/	/bakiþ/	by analogy with the infinitive
6	/bakan/	/bækiþ/	i-mutation
7	bacan	bæcþ	after further changes, irrelevant here

Analogy took place between related forms of a single lexical item, e.g. different forms of the same verb or noun. It generally did not take place between related lexical items derived from the same root, e.g. between talu "tale" and tellan "to tell".

2 This entry is misleading. Back mutation actually took place after i-mutation; this is why the result of applying both i-mutation and back mutation to a is eo rather than ie, the normal i-mutation of ea. Note also that back mutation applies only when the following syllable contains a, o, u, while i-mutation applies only when the following syllable contains i, j; hence you would not expect both back mutation and i-mutation to apply in a single word. All instances in which this occurs had one suffix substituted for another between the operation of the two processes. For example:

• Latin asellum "donkey" > Proto-Germanic *asilu (replacement of Latin diminutive suffix -ell- with similar Proto-Germanic diminutive suffix -il) > *æsil (a-fronting) > *esil (i-mutation) > *esel (a normal change in unstressed syllables) > esol (substitution of more common -ol for less common -el) > eosol (back mutation)
• Proto-Germanic *awī "ewe" > *awi (vowel reduction in unstressed syllables) > *ewi (i-mutation) > ewu (feminine -i disappeared in prehistoric Old English and was replaced with -u; a similar change occurred in e.g. menigu "multitude", cf. Gothic managei /managī/) > eowu (back mutation)

3 Proto-Indo-European /e/ was already mutated to /i/ in Proto-Germanic in two contexts: When occurring before /n/ plus consonant, and when occurring before /i/ or /j/. The more general i-mutation that applied to all vowels in Old English is a separate process that occurred many centuries later, although it had the same effect on /e/. (Note that due to this earlier change there were few instances of /e/ that could be affected by Old English i-mutation. For this reason, the i-mutations of /e/ are listed in parens, e.g. (i), to indicate that the given results are not due directly to i-mutation of /e/, but to i-mutation of /i/ or of some vowel derived from it, e.g. io.) This is also why the Proto-West-Germanic form of hierde "shepherd" appears already as *hirdija(z) with /i/ in the root even though it's clearly related to heord "herd" (Proto-West-Germanic *herdō). It's also why there's no entry for "+nf,nþ,ns" under /e/ even though it occurs for all other vowels. Furthermore, describing i as the i-mutation of e, or ie as the i-mutation of eo, is misleading at best. In fact, as just described, e was not mutated to i by i-mutation, but rather in an i-mutation environment i already appeared due to the earlier mutation of /e/ to /i/. Similarly, eo from earlier /e/ in a "breaking" environment was not mutated to ie by i-mutation. In this case again, /i/ already appeared in the i-mutation environment, which was broken to io due to the "breaking" environment it was in, and this io was then mutated to ie by i-mutation. Note further that the breaking environments for /i/ were more restrictive than those for /e/. Hence it's possible for post-breaking non-umlaut-context eo to correspond to umlaut-context i rather than io (e.g. before lh or lc), and therefore for a post-umlaut alternation between eo and i to exist. Presumably, these anomalous alternations were mostly eliminated by analogy.

4 A very similar process to what's described in note 1 resulted in the umlaut of /o/ sometimes appearing as /y/ (the "normal" outcome), and sometimes as /e/ (by analogy). Just like a-restoration, a-mutation (which lowered /u/ to /o/ before /a, e, o/) was blocked by a following /i/ or /j/, and the /u/ that was left over was sometimes changed into /o/ by analogy, and sometimes not changed.

5 Proto-Germanic mutation of /e/ to /i/ before /i/ or /j/ also affected /eu/, producing /iu/. In fact, /iu/ occurs only before /i/ or /j/ in the following syllable, and /eu/ never occurs in these circumstances. That is, /iu/ is in fact an allophone of /eu/. It is typically written as /iu/, rather than, because in the later Germanic dialects the reflexes of the sound do in fact become separate phonemes.