Jump to ContentJump to Main Navigation
Show Summary Details
More options …

Folia Linguistica

Acta Societatis Linguisticae Europaeae

Editor-in-Chief: Fischer, Olga / Norde, Muriel

Folia Linguistica
IMPACT FACTOR 2018: 0.463
5-year IMPACT FACTOR: 0.647

CiteScore 2018: 0.59

SCImago Journal Rank (SJR) 2018: 0.284
Source Normalized Impact per Paper (SNIP) 2018: 0.971

See all formats and pricing
More options …
Volume 40, Issue 1


The history of /-n/ loss in English: Phonotactic change with lexical and grammatical specificity

Donka Minkova / Michael Lefkowitz
Published Online: 2019-07-30 | DOI: https://doi.org/10.1515/flih-2019-0010


In Old English, /-n/ loss started in early Northumbrian and spread to the southern dialects after about 1050. An important diagnostic of the transition to Middle English, the loss is commonly assumed to be morphologically driven. However, /-n/ loss in atonic syllables could also be phonologically-conditioned: aweġ ‘away’<onweġ, abūtan ‘about’<onbūtan. In Middle English, the loss proceeded rapidly, but the triggers behind the different rates of change and the different results for the various categories have not been fully explored. Using LAEME, we survey all attestations of /-n/ loss, enriching the empirical data-base on the change. The findings show significant differences within word-classes, and differences between inflectional and derivational suffixes. This raises a set of theoretical questions: why did only /-n/ inflections lose their codas, why was the productivity of verbal derivational /-n/ phonotactically restricted, what justifies the loss or retention of /-n/ in stems? We look into the interplay of phonological and morphological factors, isolate the sets in which the results appear to be phonotactically driven, and address the phonotactic dimension in relation to other factors, both within and above the word level. In noun plurals, /-n/ loss emerges as the clearest case of avoidance of phonotactically suboptimal sequences at the word level. A statistical comparison of the end-points of the change reveals that overall frequency has stayed constant and has no obvious direct bearing on the process, while the presence of /-n/ as a morphological marker has changed significantly. The paper ends by identifying aspects of the history of /-n/ that remain uncharted.

Keywords: nasal loss; coda clusters; syncope; sonority sequencing; stem faithfulness

1 Introduction

The alveolar nasal /n/ in weak syllable codas 1 is one of the casualties in the erosion of the English inflectional system. It is the only singleton-coda stop in the history of English subject to deletion. 2 This contribution surveys the recoverable evidence for /-n/ changes in various morphological classes, and describes how phonotactic, lexical and grammatical factors interacted to produce the attested developments.

Final /n/-loss never affected the entire lexicon evenly; both stem-final and derivational /-n/ has been preserved in some items, as Table 1 shows.

Table 1:

/-n/ loss, replacement, or preservation in diachronic English.

Identifying the factors for the variability of final /n/-loss presents a challenge: already in late Old English it cannot be characterized as either purely phonological or purely morphological (see already Moore 1927). It raises the following more specific questions. Why did only /-n/ inflectional suffixes lose their codas? Why was the productivity of verbal derivational /-n/ phonotactically restricted to obstruent-final monosyllables (blacken, harden vs. *greenen, *soliden)? What conditioned the loss or retention of /-n/ in stems? In what sets was final /n/-loss (primarily) phonotactically driven?

To address such questions, we first discuss developments in Old English and other early Germanic languages (Section 2). In Section 3 we use the Linguistic Atlas of Early Middle English (LAEME) to extend the limited data-base previously available (specifically Moore’s twelfth/thirteenth-century data). On the basis of LAEME data, we survey the inflectional, derivational, and stem-final attestations of /-n/ in Middle English, identifying leaders and laggards in the process, focusing on differences among and within word classes, and on differences between inflection and derivation. In Section 4, we discuss the phonotactic dimension of /-n/ loss in relation to other factors, both within and above the word level. We show that /-n/ loss in noun plurals was the clearest case of avoidance of phonotactically suboptimal word forms. In Section 5, we discuss Present Day English data, which reveal that the overall frequency of final /-n/ has stayed constant, while its use as a morphological marker has changed significantly. In Section 6, we summarize our findings, present our conclusions, and identify aspects of the history of /-n/ that remain unclear.

2 /-n/ loss in Old English and beyond

2.1 /-n/ loss in the Germanic family

The loss of /-n/ in unstressed syllables is a recurrent Germanic phenomenon. Proto-Germanic (PrG) final /-n/ reflects both Indo-European /-m/ and /-n/, and was already unstable during the transition from IE to PrG (Reszkiewicz 1973: 99; Ringe 2006: 85–87; Fulk 2018: 120–121). Final /-n/ lost in Proto-Germanic fails to show up in Old English, as the examples in (1) illustrate.

IE *sunu-m>PrG *sunu(n)>OE sunu ‘son’
IE *urdho-m>PrG *wurda(n)>OE word ‘word’
IE *edono-m>PrG *etana(n)>OE etan ‘eat’
IE *tā-m>PrG *þōn, þō>OE þā, þa ‘the, that’

In Proto-Nordic (pre-800), /-n/ was lost in unstressed syllables (Nielsen 1981: 248–249) in conjunction with the nasalization of the preceding vowel: Proto-Nordic *valjan ‘to choose’>velja or veljã (Riad 1999: 136–140). In Old Norse and Old Frisian, inflectional /-n/ was lost throughout (Robinson 1992: 192), while in Old High German, Old Saxon, and most Old English varieties, the nasal was preserved. 3 Also, /-n/ was the only singleton coda consonant that could be lost even in monosyllables, although only when the preceding vowel was long. This testifies to the strictness of the minimal word constraint in early Germanic and suggests that Germanic /-n/ loss had a phonological dimension.

2.2 /-n/ frequency in Old English

In Present Day English, /n/ is the most frequent consonant overall, and in final position it is second only to /-t/. Therefore, its frequency prior to Middle English may be relevant. A search for headwords ending in <-n> in the online Bosworth-Toller Anglo-Saxon Dictionary (ASD) 4 returns nearly 17,000 hits, exceeding other final letters by far (8870 words end in <-e>, 4076 in <-s>; 3288 in <-t>; 1835 in <-r>; 1184 in <-m>). This is confirmed by tentative counts in the Dictionary of Old English Web Corpus (DOEC), 5 based on tokens in running text. 6 In the verse section, where the proportion of loans is extremely low, 27,310 tokens (about 15% of all words) end in <-n>. Thus, final /-n/ was about as frequent in Old English as in Present Day English.

However, the word class distribution among Old English <n>-final word-forms is lopsided (Figure 1): the dominant word-class entry is nouns, yet about 80% of <n>-final dictionary entries are verbal infinitives (10,413 verbs, 1715 nouns, 644 adjectives, 162 adverbs, 14 pronouns). Although these numbers need to be treated as approximations, 7 they make it fairly clear that inflectional, or suffixal, final /-n/ accounts for clearly more than half of the occurrences of <-n>.

Distributional observations on Old English from the Bosworth-Toller Anglo-Saxon Dictionary.
Figure 1:

Distributional observations on Old English from the Bosworth-Toller Anglo-Saxon Dictionary.

It is not surprising therefore that /-n/ loss is commonly seen as “morphologically driven” (Hogg 1992; Kitson 1992), without specifying the type of morphology involved, inflectional or derivational. On the other hand, the phonological side of the change is treated as primary in the Corpus of Narrative Etymologies (CoNE), where Final Nasal Deletion ((FND)) is classified as a phonological change with a caveat: “This deletion is not always phonological, but may be morphologically conditioned”. 8 As we argue in Section 4, a blanket morphological account falls short of covering all manifestations of /-n/ loss.

2.3 The Northumbrian story

In Northumbrian, Old English /-n/ loss was particularly wide-spread (Sievers 1896: 98; Wright and Wright 1925: § 288; Campbell 1959: 189–190; Hogg 1992: 305; Fernández Cuesta et al. 2008). It is found in:

  1. Weak nouns: foldu ‘earth’, cofa ‘cave’, eorðu ‘earth’, acc.sg; galla ‘gall’ dat.sg

  2. Infinitives: herġa ‘to praise’, arīsa ‘to arise’

  3. Verb pst.pl and sbjV.pl: bismærædu ‘they reviled’, cwomu ‘they came’ fæsto ‘they made fast’; geworde ‘they may become’; onfunde ‘they may have discovered’

  4. Adverbs: biġeonda ‘beyond’, binna~binne ‘within’ ūta ‘out’, befa ‘before’

  5. Numerals: twoege ‘twain’; seofo ‘seven’, tea ‘ten’

The fact that /-n/ loss occurred in diverse lexical categories like adverbs and numerals (but not, as Hogg 1992 points out, in past participles) suggests that there was a phonological factor at work. Hogg (1992: 305) also points to variation within categories and adds that it “may imply” nasalization of the preceding unstressed vowel, in which case the orthographic <-n> might stand for a nasalized vowel, as in Old Norse (see Section 2.1). Since nasalization is more likely to affect long and stressed vowels (Krakow 1994; Hajek and Maeda 2000), however, 9 one should be cautious when adducing it to account for Northumbrian Old English, where it affected short vowels in unstressed suffixes (see also Section 6 below).

3 /-n/ loss and preservation in Middle English

3.1 Overview

Loss of /-n/ is clearly related to loss of inflections and to syncretism: out of 13 different inflectional suffixes ending in /-n/ only the strong past participle -en (PDE giv-en) survives. Derivational -en fares a little better: although the feminine suffix (OE gyden ‘goddess’), and the adverbial ‘direction-from’ suffixes (OE beforan ‘before’, hindan ‘from behind’; Campbell (1959: 280); Hogg and Fulk (2011: 183)) were lost, adjectival -en (golden, silken) is still around and has been quite productive after 1400. Stem-final /-n/ is vulnerable in clitics (away<OE on-weg, aright < OE on-riht; ME i waterr ‘in water’, i bandess ‘in chains’), and is involved in the a/an allomorphy. Sporadically, stem-final /-n/ is also lost, as in eve-even, beck-beckon, holly, ivy. 10 Stem-final nasal–nasal and liquid–nasal clusters can also lose the [-n]: damn, solemn, condemn, column, 11 ell<OE eln, AmE /kɪl/ for kiln.

The metaphorical mass grave in Table 2 shows syncretism and morphological analogy strongly involved in /-n/ loss, yet it obscures finer distinctions in the way in which it unfolded, and raises questions about the potential role of phonological factors. The empirical questions are: Did inflectional /-n/ loss progress in tandem across word classes? Was /-n/ loss within individual word classes more advanced in some parts of the paradigms? Was derivational /-n/ equally productive in Old English, Middle English, and Present Day English? And on the theoretical level, the following questions are still open: Why did inflectional suffixes with other final consonants not lose their codas as well? Why was the productivity of verbal derivational /-n/ phonotactically restricted? Why did most stems in /-n/ retain their coda? Why did a few lose it?

Table 2:

Diachronic loss of /-n/ (including [-n̩] co-varying with [-ən]).

3.2 Inflectional /-n/

Inflectional /-n/ loss is attested in the southern dialects after about 1050 (Moore 1927; Kitson 1992). Table 3 summarizes its status in verbal paradigms.

Table 3:

Verbal inflectional -n.

3.2.1 Inflectional /-n/ in verbs

Although /-n/ did not survive in either infinitives or plurals, its loss progressed at different rates in the two forms. Figure 2 (based on counts from Lass 1992: 97–98) compares the rates of /-n/ loss in infinitives and in present plurals.

Retention of inflectional /-n/ in infinitives and verb plurals (PC = The Peterborough Chronicle).
Figure 2:

Retention of inflectional /-n/ in infinitives and verb plurals (PC = The Peterborough Chronicle).

Infinitival /-n/-marking was completely lost by the third quarter of the fifteenth century, while present plurals lost /-n/ more slowly and it still occurred in Early Modern English. A similar pattern was revealed in LAEME data, where we compared the progress of /-n/ loss in infinitives to its progress in past tense plurals. Our results are shown in Table 4 and Figure 3.

Loss of /-n/ in infinitives and past plurals in LAEME from c. 1250 to c. 1350.
Figure 3:

Loss of /-n/ in infinitives and past plurals in LAEME from c. 1250 to c. 1350.

Table 4:

Loss of /-n/ in infinitives and past plurals in LAEME from c. 1250 to c. 1350.

Once again, infinitives lead the way, with past plurals following. The most rapid demise of final /-n/ occurs roughly after 1250 until the end of the covered period, c. 1325. It is in the second half of the thirteenth century that the greatest gap in the two sets of data occurs.

Both the LAEME data and the Lass counts (Figure 2) suggest that /-n/ loss could progress at different rates in different paradigmatic forms, although it is difficult to identify the reason for that. We hypothesized that there might be a correlation between the loss of /-n/ in infinitives and the rise of marked infinitives, i.e. infinitives preceded by to, til, for. However, a detailed study of infinitive forms in LAEME data revealed no such correlation. In fact, final /-n/ was proportionally more frequent in marked infinitives than in bare ones.

A confounding – but potentially interesting – factor in our analysis was that the “distinctively Midland” (Lass 1992: 137) present indicative plural /-n/ was a Middle English innovation, while in subjunctive and past tense plurals (which we studied in LAEME) -en was an inherited suffix on its way out. What is interesting about this is that a form that was disappearing, i.e. the past tense plural marker, could at the same time gain in frequency by appearing in present tense forms as well (Table 5). A rough aggregate of over 40% <n>-final present plurals in LAEME shows how well -en had come to be established in that new function, at least in the Midlands.

Table 5:

Middle English present plural formation (LAEME totals).

Another phenomenon that is possibly related to the spread of present plural /-n/ in Middle English is the appearance of historically unjustified <-n> in subjunctive singulars. According to CoNE “[t]his phenomenon is widespread but rare (it appears in the present in 22 texts in LAEME CTT and in the past in 7), but with wide geographical distribution, in texts from both east and west”. 12 A possible interpretation of this development is that past tense plural -en had lost tense-specificity, and could therefore spread to other plurals. Loss of (morpho-)semantic specificity is of course known to go hand in hand with phonological erosion.

3.2.2 Inflectional /-n/ in nouns

The nasal suffix was also well represented in Old English weak nouns and adjectives. Its loss in nouns (Table 6), was mostly brought about by morphological analogy.

Table 6:

Nasal-final noun inflections.

The loss of the singular dative marking was clearly a morphosyntactic change, part of the general Middle English loss of case distinctions.

In the genitive singular, and in all plural forms, /n/-final suffixes all fell victim to analogical extension of the more frequent sibilant-final inflection, although weak nouns (-an declension) amounted to a healthy 10–15% of nouns in Old English (Hogg and Fulk 2011: 12, 50). 13

Unlike singular -en, plural -en shows some features worth looking into. Plural marking has been the most salient morphological feature in the nominal paradigm throughout the history of English, and weak nominal plural -(e)n was not only maintained in southern dialects of Middle English, but was even extended to nouns which had not had /-n/ plurals in Old English. Roedler (1911), Mossé (1952: 50), Britton (1959), Mincoff (1967: 264–267), and Wełna (1996: 88–91) provide lists of innovative Middle English -en plurals. Noticeably, plural -(e)n is attested also in Old French/Anglo-Norman loans (examples in Roedler 1911: 482–485), as Table 7 illustrates.

Table 7:

Innovative -en noun plurals in Middle English.

In Present Day English, noun -en plurals are still frequent Southern and South Midland dialects: “In s.w. dialect, chick is singular, chicken plural” (OED, s.v. chicken). Wakelin (1972: 109–112) reports 30 different -en plurals; there are also double plurals such as hipsen ‘hips. The productivity of -en as a noun plural in non-northern dialects may possibly also have supported its generalization as a verbal plural maker as well, and this may help explain its extension from the verbal past tense forms to present tense forms, although this hypothesis is difficult to test and needs to remain tentative. 14

3.2.3 Inflectional /-n/ in adjectives

Among adjectives the loss of the inflectional nasal has been carried through completely except for the now archaic possessives mine, thine.

In Old English, all case/number forms (except nom.sg and gen/dat.pl) in the weak paradigm had /-ən/; that homophony compromised the entire morphological marking system. Moore (1927: 255) identifies the very advanced /-n/ loss in the adjectival paradigms in the eleventh and twelveth centuries as “the result of sound change alone”. By early Middle English most weak adjectival inflections were reduced to -e or zero.

Despite a broader pattern of syncope in other syntactic functions, the syllabicity of weak inflections was retained in the determiner + adjective + noun frame as a rhythmic buffer, rather than as an inflectional marking: weak adjectives were always pre-nominal, and since 82% of nouns were consonant-initial (Minkova 1991: 177–185), final /-ə/ was preferred to final /-(ə)n/, as Table 8 illustrates. We return to the phrase-internal phonotactics of such strings in Section 4.1.

Table 8:

Weak adjectival inflections in the determiner + adjective + noun frame.

3.3 Derivational and stem-final /-n/

In Present Day English, non-inflectional /-n/ is evenly distributed between derivational and stem-final occurrences, and the bulk of the derivational occurrences are borrowed suffixes such as -ion, -ine. This is consistent with the expectation that inflectional and derivational suffixes are treated differently in the grammar, and that derivational suffixes are more word-like. It is also consistent with research on homophonous affixes such as /-s/, showing that morphemic status can affect the phonetic duration of word-final segments below the level of awareness (see Plag et al. 2017, and references there).

Of the four Old English derivational suffixes in <-en>, only adjectival and verbal -en survive in Present Day English, as in ashen, hearten (see Table 2). The productivity of these suffixes shows a U-shaped diachronic trajectory. As to adjectives derived by -en,

of the many words of this formation which existed in Old English scarcely any survive in modern use; but the suffix was extensively applied in Middle English to form new derivatives. […] In s.w. dialects […] the suffix [< -e(n)>] is of common occurrence, being added without restriction to all [nouns] denoting the material of which anything is composed, as in glassen, steelen, tinnen, papern, etc. (OED).

While -en adjectives are a well-defined semantic subset, we are not aware of word-level phonotactic constraints on their productivity, so along with silken, wooden, we find ryen, treen, leathern. 15

The verbal /-n/ suffix has also been stable and productive in post-Middle English; its semantic range is limited to causative or inchoative verbs (Miller 2014: 32–33). Phonotactic restrictions on new -en verbs have been recognized for a long time: their stem is preferably obstruent-final, licensing deepen, harden, soften, threaten, darken, loosen, redden, but avoiding *greenen, *warmen, *coolen, *clearen (Jespersen 1942: 355–358; Marchand 1969: 272). 16 Also, adjectival and verbal -en derivation is restricted to monosyllabic, mostly native stems: wooden, shorten, but not *calciumen, *elegancen. All surviving and newly formed -en adjectives and verbs are maximally disyllabic. 17

Stem-finally, the behavior of /-n/ depends on prosodic prominence, syllabic structure, and coda phonotactics. Some Old English and Middle English proclitics lost the nasal before consonant-initial hosts: OE on->ME a-, as in again, afore, among, afloat, alive. The loss is most regular in adverbial phrases: aback, afar, anew, asunder; around; /-n/ loss is also the first step in the development of the progressive form: on + -ing>a + -ing>-ing: on singing>a-singing>singing. Prosodic weakness is the motivation behind the loss of /-n/ in pronominal ME man~me ‘one’. 18 The best evidence of phonotactic conditioning of /-n/ loss in proclitics are the indefinite article Old English ān ‘one’, ME an~a, and the possessives OE mīn, ME min(e)~mi ‘my, mine’, OE þīn, ME thin(e)~thi ‘thy, thine’. In these well-studied cases of inherited allomorphy the preservation of the nasal, rather than its loss, serves to improve clitic-group-internal phonotactics (Schlüter 2005, Schlüter 2009, this issue).

Major class lexical items generally preserve their final consonants. There are no monosyllabic -Vn# lexical items in which /-n/ loss is recorded. The diachronic picture in Table 2 matches synchronically observable facts. Current phonetic research (Plag et al. 2017) shows that phonologically identical segments are perceptibly longer if they are stem-final than if they are affixal. 19 In the few items like holly, ivy, game, eve, lent, oft /-n/ loss has affected the weak syllable coda of disyllabic inputs, in some cases producing a lexical split, as in even(ing)~eve, often~oft, and corroborating the link between the morphological and prosodic status of /-n/ and its acoustic prominence. Further, the realization of /-n/ is also blocked if it is preceded by another nasal (see Table 2). This instantiates the sonority sequencing principle, privileging falling sonority in coda clusters. In liquid–nasal codas, outcomes vary: historical /-n/ is lost in ell, mill, toll, varies in kiln [kɪl ~ kɪln], and is preserved in Lincoln [‘lɪŋkən].

Phonological factors in /-n/ loss are related to the Middle English process of nunnation, “the addition of an ‘inorganic’ (etymologically unsupported) n at the end of a word […] not related to analogical plural” (CoNE; see also McShea 1933). The very existence of the process suggests that the high frequency of /-n/ affected its informativity. Also, if the trigger of nunnation is not morphological analogy, it indicates an ongoing /n/#~Ø# variation, not unlike the /r/#~Ø# variation attested in late Middle English (Minkova 2014: 125–126). Unlike coda /-r/ loss in non-rhotic varieties, however, coda /-n/ loss was never fully implemented, with phonological /-n/ loss being further constrained by morphosyntax.

4 The phonotactic dimension

Although the bulk of the Old English data can be handled in terms of morphological analogy and leveling, some aspects of /-n/ loss are at variance with a strictly morphological account. The Northumbrian /-n/ loss in a range of lexical categories including adverbs and numerals is at least partly non-morphological, and so is the behavior of /-n/ in verbal derivation and stems. Possible phonotactic conditioning of inflectional /-n/ loss also merits attention.

4.1 Phonotactics above the word level

The pre-vocalic vs. pre-consonantal rates of loss and retention of /-n/ in Northumbrian are not very different. Tenth/eleventh-century textual material is sparse. A spot-check in the DOEC of the context for seofo ‘seven’ shows /-n/ loss 13 times before consonantal onsets, 2 times before stressed vowels, and none before unstressed vowels. 14 out of 49 occurences of twoege ‘two’, are before vowel-initial clitics, but 11 of these are in, on, and, making it likely that what was avoided was a clitic-group-internal /-n/ haplology. A very detailed examination of the Northumbrian forms (Fernández Cuesta et al. 2008: 140) does not yield much information beyond the observation that the final nasal loss is “more general” in the later material. 20 The paucity and inconsistency of the data discourage further phonotactic inquiry.

Phonology may have been involved in loss of /-n/ in verbs (see Section 3.2.1) because the loss was most frequent in plural verbs preceding subject pronouns. For Old English, CoNE defines a Post-Pronoun Adjacency Rule ((PPAR)) 21 which includes /-n/-less forms derived from reduction of a consonant cluster, either in <-en>+ ge/we (Hogg 1992: 297, 305) or in <-an>+ ge/we, or both (see also Sievers 1896: 98, 196; Schendl 1996: 149; Benskin 2011 for Middle English). One Middle English text that we explored in detail, Havelok, has practically no <-n> final plural verbs before subject pronouns, confirming Reed (1950: 273–274, 277) and CoNE’s statement that “[i]n LAEME, the loss of /-n/ in plural forms followed by a subject pronoun is regular”. Clitic-group-internal avoidance of a consonant cluster */-nw-, -nj-, -nθ-/ would indeed favour /-n/ loss in plural verbs. 22 Once again, however, the empirical basis is limited: Middle English subject pronouns are rarely post-positioned in declarative clauses. We therefore suggest that the role of phonotactics in the case of verbal inflectional /-n/ was ancillary, and not sufficiently strong to impede the spread of /-n/ into present plurals in Middle English (see Table 5).

The avoidance of phrase-internal /-n/#C clusters in weak Adj. + N phrases (see Section 3.2.3; Table 8) is another possible phonotactic link in the course of the change. Thus, an Old English phrase þam godan burge ‘the good town’ (LS 29 (Nicholas)) takes the Middle English form þe gode boru (Havelok 773). Considering the disproportionately high frequency of consonant-initial nouns in Old English (of 22,710 nouns only 4071 are vowel-initial, i.e. under 18%), 23 the use of /-n/ in weak adjectives before consonantal onsets would have produced (dispreferred) consonant clusters over 80% of the time, while its loss would have no phonotactically undesirable effects.

In principle, a word boundary reduces the probability of phonotactic interaction between the segments flanking it. Yet, the very advanced loss of the nasal in the historically weak adjectives, combined with the preservation of the syllabicity of the inflection (see Section 3.2.3), increases the probability that phonotactic conditioning may have played a role in the development of adjectival inflections, although this hypothesis needs to be tested against quantifiable data.

4.2 Word-level phonotactics

As noted in Section 3.2.2, /-n/ loss in the singular forms of weak nouns was early and complete, while weak /-n/ plurals were more entrenched. In Southern Middle English, they even spread to other declensional classes. The viability of /-n/ in the nom/acc.pl of non-weak nouns in early Middle English was possibly enhanced by dative (OE -um> ME -en) and genitive (OE -ena> ME -en) plurals. Variability of the plural noun suffix in nouns not belonging to the Old English -an declension continued into the fifteenth century and is still identified as a regional option in Present Day English. Still, forms such as honden, horsen, housen did not survive, making the sibilant plural the norm in Present Day English.

The widely-accepted explanation of the loss of plural /-n/ in nouns is analogical spread of the more frequent sibilant plural (Mossé 1952: 48–49; Paddock 1989; Lass 1992; Fulk 2012: 57). 24 Attempts to identify phonological conditions for this process, have not been very successful. Moore (1927) examines inflectional /-n/ loss in relation to the following segment and does not identify a statistically significant phonotactic interaction above the word level. Following up on Roedler’s (1912–1916) hypothesis, based on very limited data, that /-n/ plural allomorphs in the southern dialects are preferentially attached to vowel-final stems, Newman (2009: 28) surveys the distribution of /-n/ vs. /-s/ noun plurals and concludes that “estimating the extent to which this conditioning operated in these dialects in either the twelfth century or the thirteenth century is problematic”.

We believe that a closer inspection of word-internal stem-affix phonotactics offers additional insights into the reasons for the ultimate dominance of the sibilant inflectional affix. Our argument would go like this: schwa syncope is routine in final unstressed [-ən] syllables, and weak nouns in Old English are ‘consonantal’, i.e. their plural forms end in a surface /C(ə)N#/ string. Schwa syncope places /-n/ immediately after a consonant-final stem and produces suboptimal final clusters, creating, for example, [-dn] in handen, [-pn] in bischopen, [-kn] in wiken in violation of the sonority sequencing principle of preferred falling sonority in the coda. 25 To repair the undesirable -VCN# either (a) the nasal suffix could be replaced with another allomorph, or (b) the nasal could be made a syllabic /-n̩/. In both scenarios, stem-faithfulness protects the pre-nasal consonant, while affix-faithfulness is weaker (Plag et al. 2017; Tanner et al. 2017). The question is why (a) wins when the allomorph that /-(ə)n/ competes with ends in an obstruent?

The key consideration comes from phonotactic restrictions on word final /-n/ in Present Day English, as seen in Figure 4.

CC# clusters in General British English (based on Cruttenden 2014).
Figure 4:

CC# clusters in General British English (based on Cruttenden 2014).

Nasal-final clusters are prohibited in unsuffixed words, with the exception of /ln/ and /rn/ (in rhotic varieties), in line with the sonority sequencing principle. On the other hand, word final /-Cs/ or /-Cz/ clusters are well formed and highly frequent in noun plurals and verbal 3sg forms, and /-Ct/ or /-Cd/ are well-formed and frequent in past tense and participle forms. Thus, when suffixes in /-n/ were in competition with suffixes in /-s/, /-z/, or /-t/, /-d/, regularizing the latter was a good strategy for avoiding ill-formed word-coda clusters.

The same phonotactic considerations can be applied to number marking in present tense verbs: 2sg -C(e)st forms and 3sg -C(e)th/-(e)s forms were both better phonotactically, and provided number marking even in the absence of explicit plural /-n/ marking in the Midlands. Similarly, the /-n/ plural in the past tense of weak verbs created suboptimal clusters /-dn/, or /-tn/, and would be less preferred than the 2sg -d(e)st Therefore, the loss of /-n/ number marking in verbs may not only have been motivated by the rise of an obligatory filled subject position and fixed SVO order in Middle English, but possibly also by a phonotactic bias against -CN codas.

The second repair option, making the final nasal syllabic, has resulted in the relatively rare and non-productive past participle /-n/ inflection: broken, eaten, driven. Along with the survivors in this set there are many instances of loss – bound, come, let, rung, sat, sung, swum – or variation: begot~begotten, shrunk~shrunken. In many cases /-n/ competes with /-d/ from the weak verbs: chided~chid~chidden, cleft~cloven~cleaved, shorn~sheared. 26 The paucity of /-n/ past participles in Present Day English (61 types in Reed (1950: 157)), 27 compared to their abundance in Old English suggests that phonotactic factors may also have played a role in their history.

Syllabification to /-n̩/ has also occurred in derivational /-n/ (see Section 3.3). Variable realization of the final syllable as [-Cən]~[-Cn]~[-Cn̩]~[-CØ] applies in this case too. The higher salience of derivational suffixes compared to homophonous inflectional suffixes (Plag et al. 2017) is in line with the mixed fate of derivational /-n/, with gender-marking /-n/ and adverb-forming /-n/ (as well as the very rare diminutive /-n/, not included in Table 2) sharing the fate of inflectional /-n/. It also helps to explain why derivational /-n/ experienced a revival in post-Middle English. Importantly, its productivity as a verbal suffix is also phonotactically circumscribed: new derivatives from non-obstruent-final stems are avoided. This local phonotactic restriction may be attributed to the insufficient sonority distance in a syllable with a sonorant onset followed by a syllabic nasal, while an obstruent onset + [n̩] is well-formed.

5 Is /-n/ special in Present Day English?

The historical uncoupling of /-n/ from grammatical marking raises questions about its status in Present Day English. In terms of overall tokens, /n/ is the most frequent consonant (12.59%), followed by /t/ (11.49%) and /s/ (7.92%) (Cruttenden 2014: 235). Word-finally, /-n/ (11.41%) is surpassed only by /-t/ (13.71%), based on the type-based entries in Muthmann (1999: 404). Thus, in spite of the sweeping Middle English changes, the frequency of word-final /-n/ has remained stable. A CELEX search for words with /-n/ categorized by morpheme type conveys a more detailed picture (Figure 5).

CELEX type counts for words with /-n/ by morpheme category.
Figure 5:

CELEX type counts for words with /-n/ by morpheme category.

The vast majority of Present Day English word types with /-n/ are either stems (of which more are polysyllabic than monosyllabic) or words with Latinate derivational suffixes such as -(e)an, -ation, -ion, -ine, -oon. Only a handful of words with native derivational suffixes survive, mostly verbal causative -en (hasten) and adjectival -en (silken). The only surviving /-n/ that is truly inflectional is in the non-productive past participle suffix -en.

The high frequency of Present Day English word-forms ending in /-n/ is clearly not due to inflectional morphology. In this respect /-n/ differs from (/-t/, /-d/), and (/-s/), the other highly frequent final consonants in Present Day English. Since CELEX also contains frequency data, the number of tokens of each morphological category can be estimated as well (Figure 6).

CELEX token counts for words with /-n/ by morpheme category.
Figure 6:

CELEX token counts for words with /-n/ by morpheme category.

In running text, the clear majority of final /-n/ occurs in function words (over 60%). Monosyllabic /n/-final stems, mostly native words, are as common as polysyllabic ones, despite making up less of the lexicon. Inflectional noun plural /-n/ gets a boost from the frequency of children, and the remaining /-n/ inflections come largely from the past participles of very frequent verbs: been, known, taken, given; been alone makes up 48% of all -en participles.

Comparing the distribution of word final /-n/ in Old English and Present Day English can be seen as a shift from final /-n/ signaling morphological complexity, to /-n/ in forms that are either simple or (typically lexicalized) derivations with latinate suffixes. This supports suggestions made in other contributions to this issue (specifically Baumann, Prömer and Ritt’s, Dziubalska- Koɫacczyk’s and Dressler’s), that word forms whose phonotactic shape signals their morphological composition are preferred. From that perspective, it makes good sense that the loss of final /-n/ in inflected forms did not extend to stems, and affected derivations only partially. The result is once again a lopsided distribution across simple and complex word forms, except that the majority relations have been reversed.

6 Summary and conclusions

This chapter has attempted to provide a descriptive account of how /-n/ loss evolved in English in its multiple manifestations: inflectional, derivational, and stem-final. We also seek to reconcile earlier positions on the nature of /-n/ loss defining the change as either primarily phonological, as in CoNE’s Final Nasal Deletion, or morphological (Hogg 1992; Kitson 1992).

Among consonant-final suffixes, only those with /-n/ underwent consonant deletion, making the change at least partly phonological. The striking lack of inflectional /-n/ in Present Day English against the wealth of Old English inflectional /-n/, however, is most commonly described as a morphological change. While this is certainly correct, the emphasis on morphological analogy obscures complex relationships on other levels. The philological detail on the demise of /-n/ in weak syllables reveals phonological and additional morphological factors and interactions that have not been registered or discussed in the literature.

The token frequency of /-n/ is extremely high both before and after its loss in Middle English. Frequency by itself is therefore non-explanatory. However, since frequency and informativity are studied together (Cohen Priva 2008, Cohen Priva 2017), the unevenness in the observed rates of loss, affecting primarily inflectional affixes, shows how the lack of informativity of /-n/ as a grammatical marker affected its diachronic development.

The empirical questions we addressed in Section 3.1 can be answered with confidence. Unsurprisingly, inflectional Middle English /-n/ loss was not progressing in tandem across word classes. The /-n/ in weak adjectives and weak singular nouns was lost particularly early. Loss of /-n/ within the same word class developed unevenly in different parts of the paradigm. The unevenness is securely attested in verbs (see Section 3.2.1) and in nouns (see Section 3.2.2). Of the four derivational types of /-n/ in Old English, two survive, and verbal derivational /-n/ displays obvious phonotactic restrictions.

The theoretical questions related to these empirical data, repeated here from Section 3.1, are: Why did inflectional suffixes with different final consonants not lose their codas? Why was the productivity of verbal derivational /-n/ phonotactically restricted? Why did most stems in /-n/ retain their coda? Why did a few lose it?

On the theoretical level, we have produced evidence that the boundary between a stem and an inflectional suffix is weakest, and allows the strongest degree of cross-boundary interaction in non-monomorphemic words. Thus, the sonority sequencing constraint selects /-z/, /-s/, /-d/ and /-t/ inflections over final /-n/ in most inflected forms. This also explains why suffixes like /-z/, /-s/, /-d/ and /-t/ were not lost like /-n/, when English inflectional morphology was radically simplified. The loss of /-n/ in noun plurals to sibilant noun plurals (see Section 4.2) is therefore not fully explainable in terms of morphological analogy, especially considering the local spread and survival of -en noun plurals in Middle English. Instead, extensive inflectional schwa syncope in Middle English would have resulted in the production of word final [-Cn] clusters, which violates the coda sonority sequencing constraint. The availability of a phonotactically better inflectional noun plural marker with a sibilant coda makes /-n/ substitution the preferred choice. The realization of a syllabic nasal [n̩] as the plural suffix might also have been possible. We suggest, tentatively, that it loses to the substitution of /-n/ by [-(ə)s] because it was a less salient exponent of plurality, though considerations such as lexical frequency, perceptual salience, dialectal distribution and standardization are also to be factored in. Sonority sequencing is a possible factor also in the loss of /-n/ plural marking in present tense verbs, where the integrity of number marking was partially preserved by the -(e)st and -(e)th/-(e)s suffixes, which are phonotactically better than /-n/.

/-n/ was lost in all productive inflectional suffixes, but it was preserved in two derivational suffixes. This mixed picture conforms to the hierarchy of boundaries: derivational /-n/ is less impacted by phonotactics than its inflectional homophone. On the question of the productivity of verbal derivational /-n̩/, we found an informative phonotactic link between the stem coda and /-n̩/. The almost categorical preference for obstruent-final stems in new forms is also accounted for by sonority sequencing – the preference for a sharp sonority rise from the onset to the nucleus.

As to the question why most stems have retained the codas, this may reflect a straightforward instantiation of a faithfulness constraint preserving full stem integrity. At the same time, we have also seen that the preservation of /-n/ in stems has led to radical reversal of the majority relations between simple and complex forms ending in /-n/. In Old English, final /n/ was typical of inflected word forms and had the potential of signaling morphological compositionality. In Present Day English, the opposite is the case, but at neither stage is final /-n/ evenly distributed among complex and simple word forms.

Why did a few stems nevertheless lose final /-n/? Faithfulness to the stem competes with positional weakness: loss in procliticized function words is therefore a highly probable outcome. The undesirability of clitic-group-internal /-nC-/ sequence is also the historical rationale for the Present Day English allomorphy in my-mine, thy-thine, and the indefinite article. In regular lexical items the (very rare) loss of /-n/ in weak syllables as in holly, ivy, eve is most likely modeled on the instability of inflectional and derivational /-n/.

The stability of /-n/ plurals in post-1250 Middle English, at least in non-northern varieties, suggests that it may have served as a generalized plural marker. This temporary grammatical umbrella function might have been a factor for the spread of /-n/ to the verbal present plural and to non-weak nouns. The more advanced loss of /-n/ in infinitives compared to the loss in verb plurals may also be attributable to the temporary entrenchment of /-n/ as a plural marker.

By way of an envoy, we recognize that there are many more areas of empirical and theoretical interest that deserve to be explored: Was French inf -en (consailen, blamen, chaungen) carried over from French 3rd person plural present tense, or was it simply assimilation to the native patterns? Was the French sibilant-final noun plural a factor in the loss of -en plural nouns in English? Why does accusative-dative syncretism favor him over hin<hine? Why did derivational -dom not change to -don? How can the interaction of morpho-logy, prosody, phonotactics lexical specificity, and frequency be modelled? How does /-n/ loss and substitution compare to /-r/ loss within the larger framework of diachronic variation and change in English?


This version of the paper has benefitted vastly from Margaret Laing’s careful critical scrutiny of our data and challenges to some of our arguments. We are also grateful for the helpful readings of the anonymous reviewers. We could not have asked for a better editorial process; all remaining infelicities are our own.


  • ASD=Bosworth, Joseph. 2010. An Anglo-Saxon dictionary online. Edited by Thomas N. Toller et al., compiled by Sean Christ & Ondřej Tichý. Faculty of Arts, Charles University Prague. http://bosworth.ff.cuni.cz/ (accessed November 2018). 

  • Benskin, Michael. 2011. Present indicative plural concord in Brittonic and Early English. Transactions of the Philological Society 109(2). 158–185. CrossrefGoogle Scholar

  • Britton, G. C. 1959. N-plurals in the nouns of Havelok the Dane. Neuphilologische Mitteilungen 60(2). 175–179. Google Scholar

  • Burzio, Luigi. 2007. Phonologically conditioned syncretism. In Fabio Montermini, Gilles Boyé & Nabil Hathout (eds.), Selected Proceedings of the 5th Décembrettes: Morphology in Toulouse, 1–19. Somerville, MA: Cascadilla Proceedings Project. Google Scholar

  • Campbell, Alistair. 1959. Old English grammar. Oxford: Clarendon Press. Google Scholar

  • CELEX=R. Harald Baayen, R. Piepenbrock & L. Gulikers. 1995. CELEX2. Web download. Philadelphia: Linguistic Data Consortium.Google Scholar

  • Cohen Priva, Uriel. 2008. Using information content to predict phone deletion. In Natasha Abner & Jason Bishop (eds.), Proceedings of the 27th West Coast conference on formal linguistics, 90–98. Somerville, MA: Cascadilla Proceedings Project. Google Scholar

  • Cohen Priva, Uriel. 2017. Informativity and the actuation of lenition. Language 93(3). 569–597. CrossrefGoogle Scholar

  • CoNE=Roger Lass, Margaret Laing, Rhona Alcorn & Keith Williamson. 2013. A corpus of narrative etymologies from Proto-Old English to Early Middle English and accompanying corpus of changes, Version 1.1. Edinburgh: The University of Edinburgh. http://www.lel.ed.ac.uk/ihd/CoNE/CoNE.html (accessed November 2018). 

  • Cruttenden, Alan. 2014. Gimson’s pronunciation of English, 8th edn. London: Routledge. Google Scholar

  • DOE=Angus Cameron, Ashley Crandell Amos, Antonette diPaolo Healey et al. 2016. Dictionary of Old English: A to I Online. Toronto: Dictionary of Old English Project. https://tapor.library.utoronto.ca/doe/ (accessed November 2018).

  • DOEC=Antonette diPaolo Healey, John Price Wilkin & Xin Xiang. 2009. Dictionary of Old English Web Corpus. Toronto: Dictionary of Old English Project. https://tapor.library.utoronto.ca/doecorpus/ (accessed November 2018). 

  • eLALME=Michael Benskin, Margaret Laing, Vasilis Karaiskos and Keith Williamson. 2013. An electronic version of a linguistic atlas of late mediaeval English. Edinburgh: The Authors and the University of Edinburgh. http://www.lel.ed.ac.uk/ihd/elalme/elalme.html (accessed November 2018).

  • Fernández Cuesta, Julia, Nieves Rodríguez Ledesma & Inmaculada Senra Silva. 2008. Towards a history of Northern English: Early and Late Northumbrian. Studia Neophilologica 80(2). 132–159. CrossrefGoogle Scholar

  • Fulk, Robert D. 2012. An introduction to Middle English: Grammar and texts. Peterborough, Ont. Buffalo, NY: Broadview Press. Google Scholar

  • Fulk, Robert D. 2018. A comparative grammar of the Early Germanic languages. Amsterdam: John Benjamins. Google Scholar

  • Giegerich, Heinz. 1999. Lexical strata in English. Cambridge: Cambridge University Press. Google Scholar

  • Hajek, John & Shinji Maeda. 2000. Investigating universals of sound change: The effect of vowel height and duration on the development of distinctive nasalization. In Michael B. Broe & Janet B. Pierrehumbert (eds.), Papers in laboratory phonology V, 52–69. Cambridge: Cambridge University Press. Google Scholar

  • Hoad, T. 1994. Old English weak genitive plural -an: Towards establishing the evidence. In M. Godden, D. Gray & T. Hoad (eds.), From Anglo-Saxon to Early Middle English: Studies presented to E.G. Stanley, 108–129. Oxford: Clarendon Press. Google Scholar

  • Hogg, Richard. 1992. A grammar of Old English. Vol. 1: Phonology. Oxford: Blackwell. Google Scholar

  • Hogg, Richard M. & Robert D. Fulk. 2011. A grammar of Old English. Vol. 2: Morphology. Oxford: Blackwell. Google Scholar

  • Hotta, Ryuichi. 2015. s-pluralization in Early Middle English and word frequency. In Juan Camilo Conde-Silvestre & Javier Calle-Martín (eds.), Approaches to Middle English: Variation, contact and change, 163–179. Frankfurt am Main: Peter Lang. Google Scholar

  • Jespersen, Otto. 1942. A modern English Grammar on historical principles. Part 6: Morphology. Copenhagen: Ejnar Munksgaard. Google Scholar

  • Kastovsky, Dieter. 1992. Semantics and vocabulary. In Richard M. Hogg (ed.), The Cambridge history of the English language. Vol 1: The beginnings to 1066, 290–408. Cambridge: Cambridge University Press. Google Scholar

  • Kitson, Peter. 1992. Old English dialects and the stages of the transition to Middle English. Folia Linguistica Historica 11(1–2). 27–87. Google Scholar

  • Krakow, Rena A. 1994. Nonsegmental influences on velum movement patterns: Syllables, sentences, stress, and speaking rate. Haskins Laboratory Status Report on Speech Research 117/118. 31–48. Google Scholar

  • Krygier, Marcin. 1994. The Disintegration of the English strong verb system. Frankfurt am Main: Peter Lang. Google Scholar

  • LAEME=Margaret Laing. 2013. A linguistic atlas of Early Middle English, 1150–1325, Version 3.2. Edinburgh: The University of Edinburgh. http://www.lel.ed.ac.uk/ihd/laeme2/laeme2.html (accessed November 2018). 

  • Lass, Roger. 1992. Phonology and morphology. In Norman Blake (ed.), The Cambridge history of the English Language. Vol. 2: 1066–1476, 23–155. Cambridge: Cambridge University Press. Google Scholar

  • Marchand, Hans. 1969. The categories and types of present-day English word-formation. München: C. H. Beck. Google Scholar

  • McShea, John J. 1933. The weak declensional and nunnated strong nouns in the Middle English Genesis and Exodus. Washington, D.C.: Catholic University of America PhD dissertation. Google Scholar

  • Miller, D. Gary. 2014. English lexicogenesis. Oxford: Oxford University Press. Google Scholar

  • Mincoff, Marco. 1967. English historical grammar. Sofia: Naouka i izkoustvo. Google Scholar

  • Minkova, Donka. 1991. The history of final vowels in English. Berlin: Mouton de Gruyter. Google Scholar

  • Minkova, Donka. 2014. A historical phonology of English. Edinburgh: Edinburgh University Press. Google Scholar

  • Moore, Samuel. 1927. Loss of final n in inflectional syllables of Middle English. Language 3(4). 232–259. CrossrefGoogle Scholar

  • Mossé, Fernand. 1952. A handbook of Middle English. Baltimore: Johns Hopkins University Press. Google Scholar

  • Muthmann, Gustav. 1999. Reverse English dictionary. Berlin: Mouton de Gruyter. Google Scholar

  • Myers, Sara M. 2014. An investigation of certain aspects of the genitive noun phrase in Middle English (1150-1500). Edinburgh: University of Edinburgh PhD dissertation. Google Scholar

  • Newman, John. 2009. Phonological conditioning and noun plural formation in Middle English. Anglica 18. 23–31. Google Scholar

  • Nielsen, Hans F. 1981. Old English and the continental Germanic languages: A survey of morphological and phonological interrelations. Innsbruck: Institut für Sprachwissenschaft der Universität Innsbruck. Google Scholar

  • OED=The Oxford English dictionary. http://www.oed.com/ (February 2019). 

  • Paddock, Harold. 1989. On explaining macrovariation in the sibilant and nasal suffixes of English. Folia Linguistica Historica 9(1). 235–270. Google Scholar

  • Plag, Ingo, Julia Homann & Gero Kunter. 2017. Homophony and morphology: The acoustics of word-final S in English. Journal of Linguistics 53(1). 181–216. CrossrefGoogle Scholar

  • Prince, Alan & Paul Smolensky. 2004. Optimality Theory: Constraint interaction in generative grammar. In John McCarthy (ed.), Optimality Theory in phonology: A reader, 3–71. Malden, MA: Blackwell. Google Scholar

  • Reed, David W. 1950. The history of inflectional N in English verbs before 1500. Berkeley: University of California Press. Google Scholar

  • Reszkiewicz, Alfred. 1973. A diachronic grammar of Old English, Vol. 1. Warszawa: Państwowe Wydawnictwo Naukowe. Google Scholar

  • Riad, Tomas. 1999. Structures in Germanic prosody. Stockholm: University of Stockholm. Google Scholar

  • Ringe, Don. 2006. A linguistic history of English, Vol. 1: From Proto-Indo-European to Proto-Germanic. Oxford: Oxford University Press. Google Scholar

  • Robinson, Orrin W. 1992. Old English and its closest relatives: A survey of the earliest Germanic languages. London: Routledge. Google Scholar

  • Roedler, Eduard. 1912–1916. (Part I); Anglia Neue Folge 28:4, 420–502 (Part II). Die Ausbreitung des s-Plurals im Englischen. Kiel: Chistian-Albrechts-Universität Kiel PhD dissertation. Google Scholar

  • Schendl, Herbert. 1996. The 3rd plural present indicative in Early Modern English: Variation and linguistic contact. In Derek Britton (ed.), English historical linguistics 1994: Papers from the 8th international conference on English historical linguistics, 143–160. Amsterdam: John Benjamins. Google Scholar

  • Schlüter, Julia. 2005. Rhythmic grammar: The influence of rhythm on grammatical variation and change in English. Berlin: Mouton DeGruyter. Google Scholar

  • Schlüter, Julia. 2009. Weak segments and syllable structure in Middle English. In Donka Minkova (ed.), Phonological weakness in English: From old to present-day English, 199–237. New York: Palgrave Macmillan. Google Scholar

  • Sievers, Eduard. 1896. An Old English Grammar. Translated and edited by Albert S. Cook. Boston: Ginn & Co. Google Scholar

  • Tanner, James, Morgan Sonderegger & Michael Wagner. 2017. Production planning and coronal stop deletion in spontaneous speech. Laboratory Phonology 8(1). https://doi.org/10.5334/labphon.96 (accessed November 2018). 

  • Wakelin, Martyn F. 1972. English dialects: An introduction. London: The Athlone Press. Google Scholar

  • Wełna, Jerzy. 1996. English historical morphology. Warszawa: Wydawnictwa Uniwersytetu Warszawskiego. Google Scholar

  • Wojtyś, Anna. 2009. Suffixal past participle marking in mediaeval English. Anglica 18. 45–68. Google Scholar

  • Wright, Joseph. 1905. The English dialect grammar. Oxford: Henry Frowde. Google Scholar

  • Wright, Joseph & Elizabeth M. Wright. 1925. Old English grammar. 3rd ed. Oxford: Oxford University Press. Google Scholar


  • 1

    Henceforth /-n/. As in PDE, optional syncope allows a singleton /-n/ coda in weak syllables to be realized as a syllabic /-n̩/. 

  • 2

    Historically there is deletion of final stops in -mb, -ng clusters, and (rarely) in -nd clusters. 

  • 3

    Campbell (1959: § 145; similarly Nielsen 1981: 249) considers final /n/-loss a post-invasion development for Old English, dating to the period immediately preceding the earliest written texts, the seventh century. 

  • 4


  • 5


  • 6

    The online counts are tentative because they do not separate Germanic from Latin items. The density of non-native words varies according to text type, see https://tapor.library.utoronto.ca/doecorpus/wordcount.html. 

  • 7

    ‘Infinitive headwords’ include synchronically inflected forms such as trymm-an ‘strengthen’, luf-i-an ‘to love’, but also derived forms such as slæpan<slæp ‘sleep’, or deopian ‘deepen’<deop ‘deep’ (see Kastovsky 1992: 383, 395–396). Thus, they are a blend of mostly inflectional and some derivational <-n> forms marked as verbs. Our counts for the other major word types are also approximations. The proportion of nouns is 50% among all word-types in the ASD, and drops down to about 13% among -final headwords, but that number is boosted by compounds in -man(n)/-mon(n). Adjectives and past participles are often duplicated, so that, e.g. ǣr-boren ‘first-born’ appears only as an adjective in the DOE, while the ASD lists it only as a participle, so they cannot be searched separately; see Hogg and Fulk (2011: 165, 171–172) on the intersection between adjectives and past participles in Old English. 

  • 8

    CoNE recognizes that “[t]here is also lexical conditioning: some words seem to resist this process, e.g. wǣpen WEAPON, berðen BURDEN, dryhten LORD, mægden MAIDEN, though these latter two items had alternate forms without -n”. 

  • 9

    The supposedly general preference for nasalized low vowels is questioned in Hajek and Maeda (2000). In Northumbrian Old English the height of vowels preceding the nasal vary: the inflection could be <-an>, but the pst.ind.pl is based on <-on/-un>. Moreover, though we only have Campbell’s (1959: § 473) word for it, and no counts, “loss of final -n extended further after ī than after other vowels”. He cites proper names in –i<-īna ‘diminutive’: Aelli, Coifi, Betti which show nasal loss. When the orthographic <-n> is missing, scribes often use non-low vowels as in: galla~galle ‘gall’; abbreviated bef’a~bef’æ~bef’e~bif’a ‘before’<beforan. The common interpretation of that practice is that the vowel in question was neutralized to some kind of schwa. 

  • 10

    OE holen, ME holin; OE ífegn~ífig, ME iven~ivi. 

  • 11

    These are early loans: column (OE), ME damn, solemn, condemn, autumn (14th c). The nasal can be restored in derived forms based on spelling, see the discussion in Giegerich (1999: 129, 131–166). 

  • 12

    CoNE, subjunctive singular -n ((SSN)). 

  • 13

    Out of 5909 gen.sg tokens LAEME shows a total of 6 tokens of -en for gen.sg This points to selective application of n-loss within the nominal paradigm, with -en forms more easily identifiable as plurals. For the Weak gen.pl the data-base is small. Hoad (1994) suggests that -an gen.pl is an extension of the -an form found elsewhere in the weak noun paradigm. In her extensive recent study of the genitive noun forms in ME, Myers (2014: 59) recognizes the possibility that “rather than being a variant of -VnV due to phonological reduction, the -Vn ending may be an ending type in its own right”. Commenting on the innovative -en for gen.pl, she suggests that “[i]t may be due to nunnation of forms which ended in -V, although phonological reduction of -VnV to -Vn is also possible” (Myers 2014: 64). 

  • 14

    Following one of our reviewer’s vigorous objection to “plurality” as a possible association between nominal and verbal /-n/ suffixation, we found a precedent for positing such an association. Discussing the factors contributing to the early (1075–1125) preservation of /-n/, Moore (1927: 256) identified “the indication of plurality” as a function shared by the weak nouns and finite verb forms. The regional distribution and the different longevity of the new noun plural /-n/ and the verb present plural /-n/ are problematic for /-n/ as a stable umbrella “plurality” marker, but that does not fully preclude a link seen in terms of loss of tense-specificity (3.2.1), extended to word-class specificity. CoNE recognizes Present Plural Transfer ((PPLT)) as a separate change, whereby /-n/ becomes “a generalised plural marker especially associated in ME with the Midlands” (CoNE). The transfer is estimated to apply to about 12.5% of the -n present plurals in LAEME. eLALME (County Dictionary: Item 421) records extension of -en, -yn noun plurals in 25 texts. 

  • 15

    If there is a phonotactic element in the preservation of /-n/ in past participles used adjectivally, it is ‘negative’, i.e. stems ending in a nasal are more prone to /-n/ loss (Wojtyś 2009). 

  • 16

    Jespersen restricts verbal /-n/ derivation to base stems with a stop or a fricative coda, repeated in subsequent coverage (Marchand 1969; Miller 2014). We formulate the restriction to obstruent-final stems because of attested richen (1795), largen (1844). The OED records only 4 very rare derivatives from non-obstruent-final stems: souren (1570), dimmen (1828), dullen (1832), shoalen (1731). The verbal suffix is fully productive for stems ending in clusters: strengthen (1405), (em)bolden (1503), dusken (1550), milden (1603), blunten (1615), dampen (1633), plumpen (1687), brisken (1799), crispen (1961), etc., the last two challenging Burzio’s (2007: 17) claim that “the English suffix -en as in sof(t)-en […] attaches only to stems that end in a single obstruent”. On the loss of the dental in moisten, chasten, but not in (rare) swiften, see Marchand (1969: 273). 

  • 17

    The disyllabic base is open to further morphological processes, e.g. deepening, foreshorten etc. 

  • 18

    See Wright (1905: 223) on dialectal loss of -n in PDE in in, on, upon, when, similar to the history of reduced on- in e.g. a-leaping, a-milking, a-swimming. “Reduced forms of in, prep., are widespread in ME and in the modern dialects; o’ is a form still found in Scottish and northern English regional usage” (OED). 

  • 19

    “We found statistically significant differences between certain types of S. […] Words with no morphological boundary before S show the longest duration of S. Words with a suffixal S have shorter S’s, and cliticized S’s are even shorter (the difference between plural and genitive, and plural and genitive-plural was marginally significant)” (Plag et al. 2017: 207–208). 

  • 20

    The OED entry on N, n. – “[…] very frequently in later texts in this dialect [Northumbrian Old English], loss of word-final -n is established before unaccented vowels […]” – must be a mistake for “[…] after unaccented vowels” (OED; our emphasis); otherwise the implied pattern of deletion would create clitic-internal hiatus, a puzzlingly marked option. 

  • 21

    “When a 1 person pl or 2 person pl pronoun immediately follows a verb in any tense or mood the verbal ending (pres indicative -aþ, -iaþ, -þ; past tense -on; present/past subjunctive -en) can be reduced to -(i)e or, in vocalic stems, -Ø: e.g. cume we WE COME, lufie ge YE LOVE, ride we WE RODE, dō ge YE DO”. See Campbell (1959: § 730); Hogg and Fulk (2011: § 6.7)”. 

  • 22

    Positing a causal relation between /-n/ loss and pronominal cliticization in plural verb forms contradicts the findings in Figures 2 and 3: while plural forms can be followed by subject pronouns, infinitives cannot. Since /-n/ loss in plurals appears to lag behind, the enclitic phonotactic dimension has a limited application and would not account for the advanced stage of /-n/ loss in infinitives. The more rapid loss of /-n/ in infinitives compared to the loss in verb plurals may instead be attributable to the temporary entrenchment of /-n/ as a plural marker. 

  • 23

    Counts from ASD: <a-> (849); <æ-> (570); <e-> (1213); <i-> (318); <o->(531); <u->(480); <y->(110). 

  • 24

    A series of studies by Hotta, most recently Hotta (2015) tracks the relationship between lexical frequency and the innovative -s pluralization; the finding that the shift to -s is led by high-frequency words up to the second half of the thirteenth century is consistent with our proposed phonotactic component of the change. 

  • 25

    On sonority distance at syllabic margins and the universal asymmetry between onsets and codas in OT see Prince and Smolensky (2004). The same rationale applies to word-final assimilations recorded in Hogg (1992: 301): hræfn~hremn ‘raven’, stæfn~stemn ‘voice’, wæpn~wæmm ‘weapon’. 

  • 26

    Hogg and Fulk (2011: 213) identify the strong verbs in Old English as a “closed class”, which would curb the productivity of the -en past participles. Krygier (1994) counts 306 exclusively strong verbs in Old English, of which only 57 remained exclusively strong verbs in the fifteenth century. Only about a quarter of the original past participles in /-n/ survive. 

  • 27

    Reed’s counts exclude compounds and words which are used only as adjectives, though they are historically past participles. Of the 61 forms only 44 are the exclusive past participles of their verbs, unlike gotten, stricken. 

About the article

Received: 2018-07-06

Revised: 2018-12-12

Accepted: 2019-01-02

Published Online: 2019-07-30

Published in Print: 2019-07-26

Citation Information: Folia Linguistica, Volume 40, Issue 1, Pages 203–230, ISSN (Online) 1614-7308, ISSN (Print) 0165-4004, DOI: https://doi.org/10.1515/flih-2019-0010.

Export Citation

© 2019 Walter de Gruyter GmbH, Berlin/Boston.Get Permission

Comments (0)

Please log in or register to comment.
Log in