¹ For the history of this change in the eleventh and twelfth centuries and for much of the evidence used in this paper see my former paper, Loss of Final n in Inflectional Syllables of Middle English, published in Language 3. 232–59 (hereafter referred to as LFn).

² These forms are:

þe and þeo for the demonstrative and definite article sē and sēo (Nos. 6, 19, 37, 38, 66, 67, 68, 77, 78, 82, 89, 103, 106)

Analogical n-plurals (Nos. 5, 24, 45, 65, 76, 80)

Use of the dative forms of the third personal pronoun instead of the accusative forms (Nos. 65, 68, 80, 83)

Analogical s-plurals (Nos. 24, 38, 65)

The numbers in parenthesis refer to the texts as listed in the Appendix. It should be understood that my statement as to the period at which these forms attained their maximum displacement of historical forms applies to the Southern dialect, not to the Midland.

³ The material I have used does not exhaust the eleventh century material that is in print but includes the great majority of prose texts available in fairly modern editions. Considerably more material is available in Liebermann's Gesetze der Angelsachsen, for I restricted myself there (somewhat unwisely, I think) to the texts composed fairly late in the Old English period. I have used only a few of the interlinear texts and have made no use of the poetical texts. Included among the eleventh century texts are a few written early in the twelfth century.

⁴ Of these 46 texts, 7 contain examples of final e in the nominative singular of feminine jō- or long feminine ō-stems and really belong, I believe, to a later period of morphological development than the other 39 texts.

⁵ These three texts are numbers 13, 34, and 77. The possibility of their showing evidence of the change of m to n is greatly restricted by their brevity, for all are less than 1000 words in length. Moreover, two other texts in the same MS (tho not in the same hand) as 13 contain examples of the change of m to n and number 76 in the same MS as 77 contains examples of the change; number 34 is the only text I have from the MS in which it occurs.

⁶ A mere inspection of the material in the Appendix is sufficient to verify this statement.

⁷ See Table I in LFn 238.

⁸ See Table II in LFn 243.

⁹ See LFn 246f.

¹⁰ These texts are numbers 4 and 92.

¹¹ In two homilies of Aelfric printed from MS Corpus Christi College Cambridge 188 by Assmann (Bibliothek der angelsächsischen prosa 3. 24–64) I found only one example of the change of m to n but five examples of -um for -an; Aelfric's Letter to Wulfsige printed from MS Corpus Christi College Cambridge 190 by Fehr (Bibliothek der angelsāchsischen prosa 9. 1ff.) contains no clear case of the change of m to n but at least two examples of -urn for -an. The other texts containing examples of final m used instead of final n are numbers 9, 10, 12, 17, 31, 37, 46, 67, 68, 71, 81, 82, 84, 89, 95. Most of the examples are um-spellings for the masculine and neuter dative singular of the weak adjective but there are also examples of infinitives and oblique forms of weak nouns spelled with final m.

¹² Of the twelfth century texts used for Table I in LFn (listed in LFn 235f.), um-spellings predominate, with some n-forms, in numbers 3, 6, 7, 8, 9, 12, 15; n-forms predominate, with some um-spellings, in numbers 1, 2, 4, 10, 11, 13; there are very few or no m-spellings in numbers 5 and 14.

¹³ The evidence of the texts that have examples of the loss of n but not of the addition of e in the feminine nominative singular is of less weight than their number might seem to indicate. Numbers 7, 27, and 58 (containing about 1000, 600, and 400 words respectively) are so short that the possibility of their showing evidence of the loss of n is considerably restricted. (Moreover, numbers 29 and 30, in the same MS and hand as 27, and numbers 55, 56, and 57, in the same MS and hand as 58, show evidence of the loss of n tho 27 and 58 do not.) With respect to the longer texts, number 48 in the same MS as 47, number 51 in the same MS and hand as 52, and number 53 in the same MS and hand as 54 show evidence of the loss of n tho 47, 52, and 54 do not. The distribution of forms in number 91 (which is partly prose and partly verse) is not paralleled in any other of the 110 texts. The verse contains 5 examples of the addition of e in the nominative singular of feminines but no example of either change of m to n or loss of n. The prose parts of the text (and also texts 89 and 90 in the same MS) contain examples of the change of m to n but no examples of the loss of n.

¹⁴ See LFn 248f.

¹⁵ The MSS referred to are those containing the texts listed in LFn 235f. The Fragments in the Worcester MS is the only text in the list that has e almost exclusively for OE a, o, and u. My statements as to the relative frequency of the spellings are only a very rough estimate based on a sample of one or two pages in each text.

^16a Number 24 is in MS Corpus Christi College Cambridge 303, the twelfth century text referred to in the preceding paragraph.

^16b I have not counted among the e spellings those that occur in present and preterit subjunctive forms and the past participles of strong verbs, which ended in en in OE.

¹⁷ The evidence is incomplete because of the indefiniteness of the criterion that those texts show evidence of levelling in which the back vowels are written e ‘with more or less frequency’. Some more objective criterion is needed, e.g. that two (or three) such spellings occur in 1000 words of text. I believe that the application of such a criterion to the texts I have used might show that there are rather more than 30 texts that show evidence of levelling. It would also make possible a crude quantitative treatment of the data that might very well show differences great enough to be significant.

¹⁸ According to Luick (Historische grammatik der englischen sprache 303, 491) and Jordan (Handbuch der mittelenglischen grammatik 128, 152) the change of m to n in the dative ending um was not the result of sound-change but of analogy. Their argument is concerned solely with showing that the change was not sound-change, not with showing that it was an analogical development. Luick's treatment of the point is more discriminating than Jordan's. For Jordan cites as part of his evidence that m ‘bleibt in allen stellungen lautgesetzlich erhalten, auch im auslaut’ the words bosom, bottom, fathom, etc., which are obviously irrelevant to the discussion because the m was protected in the final position by its retention in the inflected forms. The only words offered as evidence in Luick's discussion are Middle English hwilom and the place-name Downham (Middle English Dounum). Now hwilon is much commoner in the eleventh century texts than hwilum (there are more than 25 examples of hwilon in my Appendix) and the earliest Middle English references for hwilom in Middle English in the Oxford Dictionary are Orm, and Kentish Sermons. There is therefore good reason for accepting Jespersen's view (Modern English Grammar, 1. 2. 414) that the final m of hwilom does not come from the Old English dative plural ending but developed (like m in venom, ransom, etc.) from earlier Middle English n. Jespersen's discussion is well supplemented by Holthausen's note (Anglia Beiblatt, 31. 137f.), which points out that final m in all the words under discussion can be explained as due either to ‘assimilation at a distance’ (e.g. pilgrim) or to dissimilation in syllables that began and ended with a dental. The difficulties in the way of regarding the change of m to n as a sound-change are therefore chiefly imaginary ones. Even if there were real difficulties in the way of so regarding it they would have to be very serious indeed to justify us in attributing the change to analogy. The analogical change according to Luick was from um to un (or on), which was later weakened to an; according to Jordan it was from um to on. What analogical process could have displaced the ending um by an ending un which does not occur in Old English at all? Or by the ending on which occurs only as a plural ending of verbs? We have a right to ask that those who hold this theory shall explain what analogical process could have operated to produce this result.

¹⁹ In the proportions given here and later the analogical formation is indicated by italics.

²⁰ I wish to point out here a formal contradition between this proportion and the first proportion in LFn 253. In LFn the nominative singular form hwīle is used as one of the elements in a proportion that is intended to show that the n-less form of the dative and accusative singular hunte (and by implication tunge) was supported by the analogy of five other types of noun inflection and that this analogy tended to accelerate the loss of n. In the proportion given here the n-less forms of hunte and tunge are themselves used as elements of a proportion that is intended to account for hwīle itself. Obviously the nominative singular form hwīle was not available until it came into use. The n-less forms of hunte and tunge were available as part of the analogical material that favored the development of hwīle, however, for loss of n preceded the analogical change. And after the nominative singular form hwīle had developed it still further strengthened the analogical support that favored the n-less form of hunte and tunge.

²¹ The analogical accusative singular form dǣde (see page 249 above) did not completely displace the earlier form dǣd; both survived in late Old English. The older accusative form dǣd was one of the elements in the inflectional system that favored the development of the analogical nominative forms hwīle and synne. But on the basis of the later accusative singular dǣde there developed by analogy a new nominative singular form dǣde.

²² The words lufe, etc., represent the following types of noun inflection: short feminine ō-stems; feminine, masculine, and neuter n-stems; masculine and neuter ja-stems, masculine and feminine u-stems, feminine i-stems, masculine a-stems, short neuter a-stems, and long neuter a-stems. The very great majority of all the Old English nouns were declined according to one or other of the types of inflection included in the proportion and the nominative and accusative singular were identical in all the types not included except the feminine wō-stems, e.g. sceadu and mǣd. The development of these is complicated by certain factors that do not enter into the development of the feminine i-stems, jō-stems, and long ō-stems but in principle is the same.

²³ If the addition of e in the feminine nominative singular began before the levelling of unaccented vowels was completed the analogical process would not have had the support of the types lufe-lufe and tunge-tunge but would have had that of all the other types included in the proportion.

⁵⁰ The approximate date of the MS (if given by the editor or accessible to me in a catalog of the MSS) is indicated in parentheses after the MS notation. After (2) are cited the examples of change of m to n in each text; after (3) the examples of loss of n; after (4) the examples of the analogical final e in the nominative singular of feminine nouns. The examples of loss of n cited in my paper: Loss of Final n in Inflectional Syllables of Middle English (Language, 3. 232–59) are not repeated here but are referred to by the abbreviation LFn; e.g. the abbreviation LFn 14 (4) under text 3 indicates that the examples of loss of final n in this text are to be found under text 14 of the Appendix of that article (257ff.) and that the number of examples there given is 4. The abbreviation etc. after the last example in an exhibit shows that the text contains other examples that I have not cited; when it occurs after an example that is not the last it shows that the text contains other examples of the same form. The references are usually to page and line or to line alone. Accents and marks of quantity are not reproduced.

⁵¹ Crawford prints the text of Hatton 115 and gives the variants of the other MSS.

⁵² James, Descriptive Catalogue of MSS of Corpus Christi College, 1. 475, says as to the date of this MS merely that it is in various hands of the tenth and eleventh centuries.

⁵³ See note 51 above.

⁵⁴ All four texts appear to be in the same hand.

⁵⁵ This MS, according to James (2. 311), is all in one hand; see note 56 below.

⁵⁶ Homilies 47, 48, and 49 in this MS are in the same hand as that of CCC 416 but homily 50, pp. 266–74, is in a different hand (James, 2. 314).

⁵⁷ Omitting the passage (apparently by a different hand) which Thorpe prints in brackets and footnote, pp. 244ff.

⁵⁸ For a very full description of the MS and information as to the various hands see Förster, Archiv f. d. Studium der neueren Sprachen 121. 30ff.

⁵⁹ This is a rather questionable example of analogical final e, for seven feminine nouns in -nes follow immediately after in the same construction.

⁶⁰ I regard as a scribal error the MS me for (stressed) men in 12.

⁶¹ My data as to the language of this text are chiefly derived from Logeman's introduction, especially sections 39, 75, 78, 81, 87.

⁶² This MS is in various hands but there is not more than a few years difference between the earliest and latest, 1066; see Plummer, Two of the Saxon Chronicles Parallel 2. xxxi.

⁶³ On p. 240 (not included in the material I used) occurs herehyþe, an example of analogical final e in the feminine nominative singular.

⁶⁴ On p. 240 (not included in the material I used) occurs herehupæ, another example of analogical final e in the feminine nominative singular.

⁶⁵ This MS, written in two hands, is the first part of the Beowulf codex but entirely distinct from it in origin.

⁶⁶ This is the second part of the Beowulf codex and the hand is that of the first hand of Beowulf.

⁶⁷ The form micle given in my former paper should be deleted. It is not a weak adjective and does not show loss of n.

⁶⁸ This is not a certain example of loss of n, for heofon (usually masculine) is sometimes feminine in late OE; I assume, however, that we have here the genitive singular of the very much more frequent heofene, feminine n-stem.

⁶⁹ This MS, according to Fehr, is in several hands.

⁷⁰ All the examples of analogical final e in feminines are in the verse parts of the text and all the examples of n for m are in the prose.

⁷¹ This MS is a seventeenth century copy transcribed by William De L'Isle from a MS now lost.

⁷² Brotanek prints the text of Bibl. Nat. Lat. 943 and the variants of Lambeth 489.

⁷³ In Numbers XII: 10 appears the form hreofnysse where the other MSS have hreofnys, feminine nominative singular. But the MS is very fragmentary at this point and -nende hreofnysse are the only words legible in a space corresponding to about twenty words in Laud 509. I therefore disregard this possible example of analogical final e in the feminine nominative singular because the form may not really have been nominative in the Lincoln MS.