eME normalisation issues
issues arising in the normalisation of word forms in the East Midland dialect c 1200
This is not a page for beginners. Nor is it of relevance to those whose immediate goal is to learn Early Middle English. If you are unfamiliar with Old English, or are simply looking for concise information on the grammar of Early Middle English, I strongly recommend that you skip this page and go to the eME language section instead.
The following are some issues I encountered while attempting to normalise the East Midland dialect of PC2 and Orm, according to the criteria set out in eME normalisation of forms I.
Please note that hereinafter, the abbreviation eME refers specifically to the normalised East Midland dialect of Early Middle English which is used in englesaxe.
- unstressed penultimate syllables: there is tendency in the 13th century to reinstate the unstressed penultimate syllable; in OE, unstressed penultimate syllables were generally dropped in oblique cases before n, r and l; examples can still be found in Orm - deofell/deofless, mikkell/miccle, heoffen/heoffness, and in PC2 - od`er/od`re, castel/castles, minster/minstre, gysles; there was also a tendency in all dialects from at least the late 12th century, to insert schwa before r where it didn't previously exist - OE ae_fre > Owl euer, Lmn aeuere, AW eauer, Ch ever(e); the tendency to reinstate or insert a penultimate syllable was already evident in PC2: OE micle, strengra, ae_fre, t`e_ahhwaed`re > PC2 micele, strengere, aeuer(t), t`ohuuet`ere, and appears well established by the time of Owl (1200?); thus both options - with or without the unstressed penultimate syllable, seem acceptable in eME; however, forms such as de_fles, castles, ae_fre are preferred in this site.
- unstressed pronoun forms: e.g. man > me and ic > I_; the unstressed form me appears in PC2, Owl, Lmn and Ch, while I_ appears in both PC2 and ModE; however I've only used these in the normalised versions of PC2 texts; elsewhere in englesaxe, me is not used because it is likely to be confused with me_, while I_ is not used because ic(h) is far more common in both primary and secondary eME sources; I is only used once in PC2, where it is the only form of the 1st person pronoun recorded, while Orm has icc, and Lmn, AW and Owl have ich.
- assimilation of pt.pl. form to pt.3sg. form in class 4 and class 5 strong verbs: this process is already evident in PC2: OE g`ae_fon, bae_ron, drae_pon, stae_lon > PC2 iafen/iauen, baren, drapen, stalen, but isn't attested in Orm; however, there are counter-examples in PC2 - eten, undergaeton, braecon (note - the last is most likely subjunctive), in Orm - saey`h`enn (saw), and in Lmn - y`efven; since support for this assimilation pattern is mixed in Ch - beden/bade, beren/baren, ete(n), gatt, sete(n)/sate, speeke(n)/spake(n), yeve/yave(n), broke, I haven't applied the assimilation anywhere in englesaxe; see more on this at the Chaucer Test; it is an option however.
- vowel lengthening before consonant clusters ld, nd, ng, mb, rd: this is an unusual case in eME, in that this pattern of sound development is widely supported in Orm, Ch and ModE, but has numerous exceptions. In fact the exceptions to vowel lengthening are almost as numerous as the examples of vowel lengthening, in ME texts. (See the section on sounds for a concise description of the development.) In this regard, it seems there are four options for eME:
- retain the short vowel of OE in all cases involving ld, nd, ng, mb, rd;
- lengthen the vowel in all cases involving ld, nd, ng, mb, rd, with the sole exceptions being the counter-examples drawn from Orm: shollde, wollde, wullderr, annd, stanndenn, sennd p ptc, wenndenn pt pl, winndeclut, wunnderr, unnderr, enngell, herrde, wurrdenn pt pl;
- lengthen the vowel only in those individual cases involving ld, nd, ng, mb, rd, which are found in an eME source (PC2, Orm, Lmn, AW, Owl, SO, Ch) with clearly indicated long vowel, or which have a ModE reflex with long vowel; a provisional list of cases in which the OE word has a short vowel, but the ME reflex has a long vowel, and the ModE reflex either doesn't exist or has a short vowel, are as follows: OE beldan, held, land, hand, sand, strand, wendan, rendan, rang, sang, t`rang, wrang, king, t`ing, sungon, hierde, word > Orm beldenn, held, land /be:ld@`n/, /he:ld/, /lA`:nd/, PH3 loand, Ch hond/hand, so(o)nd, strond, Orm wendenn inf /we:nd@`n/, Ch weendeth pr sg3, reenden pr pl, rong, soong, throng, wroong, Orm king, t`ing, sungenn, hirde, word /kI`:N`/, /T``I`:N`/, /su:N`@`n/, /hI`:rd@`/, /wo:rd/, Ch woordes
- lengthen the vowel only in those cases involving ld, nd, ng, mb, rd, in which a long vowel has persisted into ModE; with the addition of the pt sg3 forms of s v3 verbs in ~inden (eg ba_nd, fa_nd, wa_nd).
- vowel lengthening in open syllables in disyllabic words: note - this type of vowel change - a lengthening of e and a, doesn't affect eME spelling since it arose in the latter half of the 13th century; however, it is employed in the late ME alternate spelling option, if supported by ModE; note - we have break & speak from OE brecan & specan, but not ModE *forgeat, *neaphew nor ME *y`ae_ven, *fory`ae_ten, *nae_ve from earlier y`efen, fory`eten, nefe.
- vowel shortening: OE co_m pt sg3 > Orm comm; so is com an additional eME form? it's difficult to determine support for this change in other ME texts since vowel length was rarely indicated; there appears to be partial support in Ch for the shorter vowel, in that the pt sg3 appears as either com or cam but never as coom, whereas the pt pl appears as both coom and com (note that Chaucer didn't always double the letter representing a long vowel); it is possible that the use of com might confuse the student of OE in which the pt sg3 form is co_m; and perhaps com is not an intuitive staging post between ModE came and OE co_m; nevertheless, on balance it appears that com should be accepted alongside co_m in eME.
- consonant doubling: there was an audible distinction in length between a single consonant and double consonant in OE. That sound distinction, was still evident in Ch (see above). In fact, OE's double consonants tend to be retained in Ch. For that reason, where OE, ME and ModE differ on the question of double consonants, I have generally preferred the OE spelling. Occasionally, where a single consonant spelling is supported by both ModE usage and either PC2, Orm or an lME example cited in Clark Hall 1, I have deviated from the OE norm and preferred a single consonant: e.g. OE sibb > eME sib > ModE sibling. But the variant is also acceptable - sibb. On the whole, it's the Chaucer principle that wins out. Thus - OE fyllde > eME fillede > ModE filled, in spite of PC2 fylde; on the other hand OE eall > eME al > ModE all, due to al in PC2, Lmn, AW and Ch. Double 'n' spellings are fairly common, e.g.: OE g`ewinn > eME iwinn, in spite of both lME i-win and ModE win. Note that it isn't always the case that a double consonant in OE becomes a single consonant in ME or ModE (or both): OE wel > ModE well. I've used wel in englesaxe. Note also that OE c`c` remains eME cc, since the related spelling cch is found in both Orm and Ch.
- shortening of prefix a_-: an unnaccented long 'a' in an OE word of two or more syllables often became shortened in ME and ModE; the most notable example of this is the prefix a_- e.g. in a_ri_sen, which became a short a- in ModE arise; another example is OE la_tte_ow > ME lattew; I have generally applied this in englesaxe;
- reduction of certain consonant clusters: at least one consonant cluster that occurred in OE leaves no trace in ME - ndg. In ME this becomes ng. This surfaced in my investigation of the reflexes of OE myndgian. I suspect this reduction was already underway in the OE period since two variants without the middle consonant - d, are recorded in the OE corpus - myngian and mynegian. As that particular consonant cluster is absent in the MED, even though Orm and PC2 have no reflexes of mydngian or its variants, I have registered the sound change indg to ing as a standard eME sound and spelling change in table X. There may be similar reductions of consonant clusters. If so, I will update these notes and table X as I come across them.
- preterite of nimen and cumen: the principle applied here is the same fundamental principle applied throughout englesaxe: the conjugation of these verbs should resemble as closely as possible those of their OE forebears except where additional eME forms are available. A little discussion is warranted however. Nominally, these 2 verbs belong to class 4; in WGmc the forms followed the regular class 4 pattern: *neman nam na_men numen, *cweman *cwam *cwa_men *cwumen; however before a nasal the WGmc <a_> became <ae_> in OE in all contexts except one - before a nasal it bacame <o_>. Similarly, Gmc <a> became OE <o> before a nasal; also Gmc <e> became <i> in OE; hence the standard OE (West Saxon) forms - niman nom no_men numen; However, nam and na_men are also found in OE texts, and the most common forms in eME appear to be nimen nam na_men numen/nomen; PC2 has numerous examples of nam and namen but none of nom or nomen for the singular and plural preterite respectively; both numen and nomen can be found in North-west Midlands c.1400 for the past participle, while Ch has nam and nomen (/O:/) for the preterite. Note that /a:/ (<a_>) had passed to /O:/ (<oa/oo/o>) by the mid 13th century. And for the past participle, Ch has ynome, the <o> of which probably represents an /u/ sound as it does in Ch come(n). Taking all this into account, the most appropriate forms for eME appear to be nimen nam na_men numen. The conjugation of cumen in eME is easier to determine. Neither OE nor eME had instances of cam or ca_men to compete with co_m co_men and com co_men. It isn't until Ch that we find cam and came.
- OE fricative <g>: OE g /G``/, remained a fricative in eME. It is represented in eME by <g> (and in the alternate eME spelling by <y`>1); However there were 2 allophones of this sound, and these developed differently in later ME. If the fricative <g> had followed a back vowel or <r> or <l> in OE, it passed to /w/: boga > boge (alt boy`e2) > lME bowe (ModE bow). If on the other hand, it had followed a front vowel in OE, it was assimilated to, and lengthened, the preceding front vowel: nigon > nigen (alt niy`en2) > lME ni_n (ModE nine). For more detail, see Fricatives & Semivowels
- OE semivowel <g>: OE g` /j/, has remained a semivowel ever since, when in the initial position (e.g. OE g`e_ar > eME y`ae_r > year). However a diphthong resulted in eME from OE /j/ preceeded by a front vowel. Thus <aeg`> became <ay`> (alt <ai>), while <ae_g`>, <eg`> and <e_g`> became <ey`> (alt <ei>). For more detail, see Fricatives & Semivowels
- OE prefix g`e-: its eME reflex i- can precede any past participle (in the absence of another prefix); this appears to be entirely optional - it is only used once in PC2; i- is only used with other parts of speech if the word (or its root) is indicated in Clark Hall 1 as being rarely found without the prefix;
- ae_ or e_: OE WS ae_ appeared as e_ in OE A under certain circumstances (where the ancestral vowel was WGmc or Lat a_). Many of these forms passed to the East Midland dialect of ME and ultimately to ModE. How do we know whether ae_ or e_ is appropriate in eME? ModE is the most practical guide here. For example: OE WS dae_d, strae_t, slae_pan, grae_dig`, nae_dl = OE A de_d, stre_t, sle_pan, gre_dig`, ne_dl > ModE deed, street, sleep, greedy, needle. On that basis we can be confident that the eME equivalents are de_d, stre_t, sle_pen, gre_diy`, ne_del. However that isn't necessarily what we find in the East and West Midland3 source texts: Orm straete dat, Lmn straet, AW strete acc/dat, PC2 daedes pl, Orm/Lmn dede, PC2 slepen, Orm slaepenn, Lmn slaepen, AW slepen, Orm clene claenesse, Lmn claene, AW cleane, Orm nedle, Orm grediy`, AW gredi. The evidence is mixed. It suggests that for many of these words, either both forms co-existed in OE Anglian, or the southern form made inroads into the Midland dialects in the Early Middle English period, or perhaps that the writers (and/or) scribes of Orm and Lmn were influenced at times by the OE WS standard. Whatever the truth of the matter is, under the principles set down in Normalisation, the standard OE WS forms with ae_ pass to eME. And the alternate (OE A) forms in e_ qualify as additional eME forms by virtue of their support in the MED (as headwords) and in ModE, and the application of the ModE test. Without the ModE test, stre_t for one, would not qualify as additional eME, since straete is found in Orm without any counter example in PC2. Of course, it makes more sense to treat the group as a whole and apply a simple rule which is easy to follow : e_ where ModE has ee. That means we have doublets in all cases except claene, since ModE has clean (not cleen). As far as the doublets are concerned, in englesaxe, I've tended to the forms supported by ModE - de_d, stre_t, sle_pen, gre_diy`, ne_del. Note also that since the equivalence - OE WS ae_ = OE A e_ where ModE has ee, is consistent, this OE > ME sound change is listed in table X. There are a couple of cases where eME has e_ rather than ae_ from OE e_a: ce_ke (OE WS c`e_ace, ModE cheek) is a result of (expected) OE A smoothing before g/h/c and appears in table X. Sce_p (OE WS sc`e_ap, ModE sheep) on the other hand, is treated as an anomaly.
- c or ch: OE c before i/e/ae was generally pronounced /tS`/. The spelling remained as c in PC2, but appeared elsewhere in ME as ch. Generally speaking in OE verb paradigms, the context determined the pronunciation. As a result, palatalisation might be present in some parts of a given verb, but not in others. Two examples are c`e_osan c`e_as curon g`ecoren and c`i_nan ca_n c`inon g`ec`inen. In late ME, the palatalisation tended to spread throughout the paradigm, regardless of context, so that the examples above became cheese(n) cheas chuse(n) ychose(n) and chyne(n) choan chine(n) ychine(n) The question is - which path should a normalised eME take? Orm has chosenn but on the other hand ycore persists until SO (and perhaps later). The MED records the preterite forms chan and chon, the first of which is found in Lmn, but also coon (from a text circa 1425). Evidently ca_n survived beyond 1200. If we were to allow the emerging ME pattern to be applied universally (a sort of extended ModE test perhaps?) in such cases, then patalisation could be spread throughout the paradigm, thus giving: ce_sen cae_s chusen ichosen and ci_nen cha_n cinen icinen. Otherwise, if we were to confine our focus to forms found in either Orm or PC2 and at least one other eME source, then cha_n would not be an additional eME form. And there would be a question mark over chusen since it is cusen that is found in PC2. I think the sensible solution is to allow palatalisation throughout any given verb paradigm. The non-palatalised forms - curen icoren ca_n are also permitted of course, as eME reflexes of the standard OE forms. Note - ce_sen and ci_nen are the only examples of such paradigms that I have found so far.
- in an earlier draft of the definition of additional eME forms, I placed conditions on the eligibility of forms which enter via the MED and ModE:
- OE (an alternate form), MED headword and ModE, but only if either:
- no examples are found in either PC2 or Orm; or
- the ModE test applies (see below)
- probable OE origin (but not recorded), MED headword and ModE, and at least one of the following - PC2, Orm, Owl, Lmn, AW, SO, Ch
- OE (an alternate form), MED headword and ModE, but only if either:
- ModE vs Chaucer: It's interesting to note that the Chaucer test achieves similar results to the ModE test in respect of the simplification (or rather - removal) of stem alternation in eME. However, there's one case in which the former overrides the latter. OE stre_a strawes > Ch stree stres and straw. Yet straw strawes is preferred to strae_ strae_s in eME, because ModE has straw straw's.
form preference: A variation on the Chaucer Test can be a useful means of choosing between rival eME forms (see section above). Put simply: if an OE form is still around in Ch, it should be used in eME. Stated with a little more finesse: if Ch uses (even if not predominantly) a form which more closely reflects the corresponding standard OE form than an additional eME form, then the eME form sourced from standard OE is preferred. Hence hit would be preferred to it, ic over I, t`encen over t`enken, man over me and swa_ over se (e.g. in hwae_r/hwa_ swa_).
That said however, I haven't always applied the Chaucer Test to variant forms. My feeling is that while it's a good idea to be clear-cut and prescriptive about grammar in any normalised language, vocabulary choice - that is the selection of one word (or form of a word) over another, should be a matter of taste. Does it matter that the usage of variant forms in eME might differ from person to person? It doesn't seem to matter in ModE. "You say either, and I say either"... Anglo-saxon without tears (in this site) is somewhat eclectic in its choice of variant forms, using predominantly but not exclusively - it, ic, t`enken, man and swa_, from the list above. I've been more consistent in my application of the "Chaucer Test" in grammar (morphology), but perhaps that's not so important either. For more, see optional grammatical variations (in eME).
- *aht?: OE a_wiht had the less common variants a_wuht, a_ht and o_ht, the last of which which probably gave rise to Orm ohht, Owl and Gaw oy`t, Ch o(u)ght and finally archaic ModE ought. OE also had a short-vowel variant aht which could explain Ch and ModE (archaic) aught and Piers auht; note - only a short /a/ in lOE would give Ch a(u)ght since long initial /a:/ in lOE would give o(u)ght). So does aht qualify as a second additional eME form alongside oht? It is a borderline case, but in the end I decided against it for the following reasons:
- since the OE prefix a_wiht tends to become short in eME, the eME form derived from the standard OE form is awiht, which would also give rise to Ch aught, if the unaccented <i> were dropped;
- the ModE aught is archaic.
- *fra or fra_: Was the <a> long or short in Orm fra? The marking of the ON vowel in frá and ModE's /ou/ in fro suggests /a:/ was the sound in Orm and PC2. This would have become /O:/ in Ch. As far as I can make out, Ch only used <fra> to indicate Northern speech, in contrast to the usual Ch form fro (representing Southern speech). That makes fra the appropriate additional eME form.
- *y`uw or y`ow: My first inclination was to allow Orm, with the apparent support of SO and Ch, to set the spelling of y`ow as *y`uw. However the pronunciation of ModE you as /ju:/ rather than /*jau/ suggests that the pronunciation of SO y`ou and Ch you/yow, was /jou/ rather than /ju:/. By extension, this would have been the prevailing pronunciation of the 2nd person plural pronoun in the southern East Midlands c 1200 AD. This makes the y`ow spelling more appropriate for eME. It also reflects the spelling and pronunciation of the standard OE form e_ow more closely. A similar argument applies to the spelling y`owre in preference to y`u_re; note also the variant OE forms i_ower/i_owre (of standard OE e_ower).
- It was difficult at first to settle on the eME additional form in cases like pot`eren and y`ugelen, especially with the former, where the putative OE form isn't given and the stressed vowel differs from MED headword to ModE. Should the eME additional form be based on the MED headword, or the ModE reflex? (Or perhaps the related MDu word?) Allowing ModE to decide the final form is a convenient solution,
- a modE test for spelling?: generally speaking, the spelling of an additional eME form is easily settled on, since the forms found in the eME sources (on which the additional eME form is based), have the same underlying pronunciation; however, should there be any doubt about the pronunciation, or a difference in vowel quality between forms found in the eME sources, then the ModE spelling should be used as a guide (if a ModE reflex exists); the only such case I have encountered so far, is y`uw/y`ow; see above; if more examples crop up, I may need to expand the ModE test at eME normalisation of forms II.
- <y`h> in previous versions of the site;
- boy`he and niy`hen in previous versions of the site;
- AW and Lmn were written in the West Midland dialect, a descendant of Mercian (one of the two OE Anglian dialects);
- Clark Hall J.R., A Concise Anglo-Saxon Dictionary, Cambridge University Press, 1960