The Purpose of Double Accenting in the Ormulum and a Possible French Connection

Abstract

Based on a study contrasting the spellings of the Ormulum’s (Oxford, Bodleian Library, MS Junius 1) Hand C with those of Orm, this article proposes that final < tt > did not necessarily indicate a short preceding vowel in the hypothesized spelling system which Orm sought to reform, and that the Ormulum’s double accent marks might serve to prophylactically counteract a spelling habit present in Orm’s house of doubling final < t > following an etymological long vowel. It argues thus against previous explanations which tend to construe the double accents as redundant markers of vowel length. Further evidence is adduced to suppose that the unexpected doubling of final < t > could have been a post-Conquest orthographical tendency arising from the intermixture of English and (Anglo-)French spelling systems.

Introduction

Orm, the author of the twelfth-century Ormulum,¹ wished by means of his collection of homilies to save the souls of the English people. As part of his efforts to achieve this, he devised a very consistent spelling system that would enable his readers to preach comprehensibly to an English audience, its most salient feature being a regular correspondence of the number of final consonant graphs with the length of a preceding vowel, illustrated by the contrast in (1).² A single final consonant indicates a long preceding vowel, but where the consonant is doubled, the preceding vowel is short. In intervocalic contexts, a double consonant graph serves not to indicate vowel duration but rather a long (i.e., geminate) consonant, enabling Orm to differentiate between sune in (2) a. and sunne in (2) b.

(1)	a.	<brid> ‘bride’ < OE brȳd	(<VC#>)3
	b.	<bridd> ‘young bird’ < OE bridd	(<VCiCi#>)
(2)	a.	<sune> ‘son’ < OE sunu	(<VCV>)
	b.	<sunne> ‘sun’ < OE sunne	(<VCiCiV>)

As a consequence, Orm cannot use his linear orthography to signal preceding vowel length in intervocalic contexts of the form < VCV > ,⁴ for which reason he sometimes adds single acutes (for a long vowel) or breves (for a short vowel) where he wishes to disambiguate (cf., e.g., Sisam, 1933, p. 8), as in (3).

(3)	a.	<wrítenn> ‘write.inf’5 < OE wrītan
	b.	<wrĭtenn> ‘write.pst.ptcp’ < OE (ge)writen

While the orthography of the Ormulum has received much attention in the literature, not all of its aspects are well understood. One aspect whose meaning has eluded scholars for some time is Orm’s use of double accent marks, described by Anderson and Britton (1999, p. 307) as a “curious and exceptional feature, which does not of itself reflect phonological contrast and appears to be without parallel in any tradition”.⁶ Double acutes are found above almost all sequences of a vowel followed by a syllable- or word-final (single) < t > . We can tell that the vowel of űt in (4) is long by the presence of a single final consonant, so what does Orm wish to tell us?

(4)	To lesenn mann kinn þurrh hiss dæþ	(P63–64)
	Űt off þe defless walde.
	‘to release mankind through his death out of the devil’s dominion’

Orm’s multi-accenting has never been discussed in detail, and “none of the explanations hitherto offered in the literature […] covers every circumstance of its use” (Anderson & Britton, 1999, p. 308). It is impossible for reasons of space to discuss exceptional cases of multi-accenting in this article, but even with regard to the regular case all previous efforts have proven unsatisfactory.⁷ I shall commence therefore with a review of such previous efforts and point out their shortcomings (Sect. "Previous Approaches"). Sects. "Audience(s)" and "Precedent" consider Orm’s audience and the precedent for double-accenting in the English and Anglo-French traditions. Having thus laid the necessary groundwork, I shall take a closer look at the spelling system Orm presumably sought to reform and conclude that double acutes might have been employed to prevent copyists from falling back into an unwelcome habit concerning the representation of final /t/ (Sect. "Hand C’s Contribution to the English Text"). I shall then draw attention to further evidence suggesting that this unwelcome habit might have arisen from the intermixture of English and (Anglo-)French spelling systems (Sect. "A French Connection?").

Previous Approaches

Deutschbein (1911, pp. 50–52) believes that Orm’s double acutes indicate a special quantity of the vowel before final < t > , positing “Halblänge” (‘half-length’) in forms like űt. He constructs an argument for a fivefold distinction of vowel quantities, which Markus (1989, p. 72) understandably finds “hardly convincing in view of the fact that […] vocalic quantity was dying out as a distinctive feature anyway”.

Murray (2002, pp. 647–648) reconsiders Deutschbein’s half-length in the light of syllable cut prosody and accredits to it “some phonetic plausibility in the sense that the apical plosives are primary sources of transition to abrupt cut”. However, the mere observation that Orm’s use of double acutes might be compatible with some decidedly modern linguistic theory should not lead us to believe that Orm was aware of the phenomenon which the theory supposedly explains. It cannot be overstated that Orm was a canon regular with very practical motivations and goals, and not an academic phonetician.

Markus (1989, p. 82–83), who summarizes and synthesizes explanations previously advanced, simply starts from the assertion that “the function of the double acutes is to indicate length in a closed syllable” and immediately jumps to the question of why Orm would have introduced this apparently redundant marking. Claiming that Orm wished to remind his readers of long pronunciation, he adduces two insights from the literature that could have led to (undesirable) shorter pronunciations before final /t/ (i.e., Trautmann, 1896/Luick 1914/1964; Sisam, 1933). There are, however, some problems with these suggestions.

Trautmann (1896, pp. 377–378; accepted by Björkman, 1913, pp. 371–372) claims that the double acutes exhort the reader not to succumb to an alleged early-thirteenth-century (sic!) tendency towards allophonic shortening of vowels before single consonants in monosyllabic words; Luick (1914/1964, §388) proposes, paradoxically, both that the alleged unwelcome tendency was strongest preceding “fortis” consonants–/t/ being one such consonant–and that Orm, because he was a subtle observer, still perceived a long vowel as the only correct choice in űt (< OE ūt ‘out’) or főt (< OE fōt ‘foot’), whereas in other words–some not ending in a “fortis” consonant–he would occasionally hear, and then spell accordingly, a short vowel. The argument is evidently inconsistent. In addition, both Trautmann and Luick fail to explain why Orm should have wished to counteract a tendency supposedly found in the dialect he sought to represent in his reformed spelling.

The second proposal referred to by Markus comes from Sisam (1933, p. 10; 1953/1962, p. 195), who remarks in a footnote that Orm’s double acutes “may be intended to counteract habits which a preacher would derive from his training in the reading of Latin”, as “it is a safe rule that all Latin vowels before a final t are short”.⁸ This statement appears to be correct for both Classical and Norman Latin, but Sisam fails to consider that the same applies–again, both for Classical and Norman Latin–to final b, d, and m as well.⁹ If Orm wished by means of his double acutes to prevent his readers from falling back into a habit from Latin speech, one would therefore expect him to be just as concerned with the frequently-occurring final − m and − d as with -t. That this was not the case is attested by the ubiquity of double acutes above final < Vt > and their simultaneous absence from < Vd > and < Vm > .¹⁰ Sisam’s idea of an encroachment of Latin vowel quantities is hence unlikely to have induced Orm’s preoccupation with final < Vt > .

In most previous studies it has either been argued or taken for granted that Orm’s double acutes have to do with vowel length; Trautmann, Luick, Sisam and Markus consider them redundant length markers. One additional and more general challenge for any such conception is that Orm in fact extended his occasional use of single acutes, illustrated above in (3) a., from intervocalic to word-final contexts in a small number of tokens to redundantly indicate vowel length (cf. Anderson & Britton, 1999, pp. 323, 325), which begs the question of why the same sign could not have been used for the same purpose preceding final < t > .

Audience(s)

To discover the purpose of Orm’s double-accenting, it is, first of all, necessary to distinguish carefully between the author’s mediate and immediate audiences (the parishioners and the readers who were to preach what Orm had written, respectively), because these groups need not have had much in common, and considering whatever distinguished them from each other may prove useful in the attempt to discover why and for whose benefit the diacritics were placed.¹¹ Orm wrote in an English largely lacking in French vocabulary (Johannesson & Cooper forthcoming: xlviii) and with the stated intent of affecting this mediate audience of ennglissh follc (‘English people’, cf. P19) by way of sermons in English, which suggests that this was their primary and probably only language. But why would preachers – most likely fellow Augustinians – have required help with pronunciation?

Orm probably lived and worked at Bourne Abbey (cf. Parkes, 1991, pp. 196–199), an Arrouaisian house founded in 1138, only five years after the first English house of the order had been established at Missenden (Robinson, 1980, pp. 59, 356). Missenden had been staffed with “canons from the abbey of St. Mary Ruisseauville, thirty miles from Arras” (Robinson, 1980, p. 59), and according to Robinson, it is likely that Bourne was in turn colonized by canons from Missenden. Worley (2003, p. 23) believes that the presumed French character of Bourne Abbey continued into Orm’s time, and her judgment has some merit: only twenty-six years lie between the abbey’s foundation and the 1164 Constitutions of Clarendon, which “had for the most part closed the priesthood to villeins–a group that included nearly all English speakers” (Worley, 2003, p. 23). In consequence, if French can justifiably be considered the language of the house in the years after its foundation, it very probably continued to be so after the doors had been shut on most monolingual English speakers.

In the second half of the twelfth century, “French became the dominant language in […] the monasteries and elite religious culture” (Faulkner, 2022, pp. 128–129); indeed, Faulkner (2022, p. 130) discerns “a growing tolerance for immigrant prelates to remain monolingual” during this period, and there is no reason to believe that Bourne diverged from the norm. A lack of competence in spoken English would have been no excuse to neglect the duty to preach, although this was clearly perceived as something of a difficulty by immigrant clerics.¹² All things considered, it does not seem unreasonable to suggest that Orm’s immediate audience of preachers required assistance with English pronunciation due to their being native speakers of French faced with the task of affecting a monolingually anglophone audience.

Precedent

Orm would have been well-positioned, as Worley (2003, p. 23) points out, to try to help improve his fellow canons’ pronunciation. If Orm wished for his sermons to be understood by all English people, he would have aimed to make it as easy as possible for a preacher to understand what was indicated by the double acutes. Whatever is conventional is easy to understand, and it is therefore necessary to look to texts found in contemporary or earlier manuscripts for writing conventions that Orm could build on and expect his readers to be familiar with.

In Old English manuscripts, precedent for double accenting seems to be scarce, and none of it is comparable to Orm’s use of double acutes.¹³ However, there is a second tradition–the Anglo-French–which suggests itself as a possible source for Orm’s double acutes. Many libraries of Augustinian houses held books in that language (for a list, see Legge, 1950, p. 111), and given what has been suggested about Orm’s circumstances, it may be supposed that he had access to texts written in Anglo-French. If that was indeed the case, he could have been familiar with other, indeed more frequently attested types of double accenting. According to Careri et al., (2011, pp. li–liii), double accents are very typical of twelfth-century Anglo-French manuscripts. They came into increased use when scribes found the Latin alphabet lacking in letters that could represent the sounds of their language, especially palatal c/g, and faded only in the thirteenth century when the problem was conclusively resolved–in the case of palatal c by the employment of the cedille or the digraph < ch > (Careri et al., 2011: li–liii). Note, in this context, that Orm adopted various other orthographic features from French either directly or via other post-Conquest English sources (cf. Anderson & Britton, 1999, pp. 304–305).

It may be useful, then, to identify possible insufficiencies in the scribal tradition which Orm attempted to reform, and to wonder why he did not think it feasible or preferable to come up with a linear solution to the problem he eventually sought to remedy by means of the double acute.

Hand C’s Contribution to the English Text

In his prolegomena, Orm dedicates eight verses (ll. 103–106) to an exhortation addressed at potential copyists, telling them to copy his words exactly as he has written them; his particular concern is with the double graphs. Johannesson (2007, pp. 116–117) suggests that bad experiences with careless copyists might have prompted Orm, anxious for his new spelling system to be transmitted correctly, to add this exhortation. “If Orm had a particular scribe in mind”, Johannesson writes, “it must have been the scribe behind ‘Hand C’” (2007, p. 117). That Hand C found it difficult (or did not try too hard) to emulate Orm’s spellings becomes apparent when one compares the orthography of text written by Hand C with text written by Orm. The following constitutes a study of Hand C’s spelling, based on their contributions on folios 43r, 62r, 67v, 69r, and 117v.¹⁴

Data and Observations

Table 1 shows Hand C’s record in conforming to Orm’s spelling conventions as regards the representation of syllable-final consonants following a vowel.¹⁵ Column one specifies the consonant of interest, column two contains forms that have a single final consonant both in Orm’s and Hand C’s performance, column three contains forms ending in a single final consonant in Orm’s text but in a double consonant in Hand C’s, column four lists forms that come with a double final consonant in both hands, and column five has forms that Orm would spell with double final consonants but in Hand C end in a single consonant.

Table 1 Hand C’s deviations from the Ormian model in the placement of final single and double consonants

It can be observed that two (or 16.67%) of what would have been 12 < VC$ > -segments in Orm’s regular spelling have an erroneously doubled consonant. Conversely, Hand C makes a single of a double consonant in 21 (or 28%) of 75 cases. While this difference cannot be considered statistically significant due to the low total of forms with–in Orm’s regular spelling–single final consonants, there are some broader observations to be made. First, Hand C successfully adheres to Orm’s spelling conventions in 64 (or roughly 74%) of 87 cases, and second, when they make a mistake it is almost always a failure to apply Orm’s innovative double spellings where customary OE spelling would have had a single consonant. This frequently happens in words where a second segment is represented correctly (e.g., < himsellf > , < underr > , < herte >). The two exceptions–to be discussed in Sect. "Unexpected Mistakes"–are < herr > (Orm. her) and < fótt > (Orm. főt).

Moreover, there is a conspicuous disparity in Hand C’s performance between mono- and multisyllabic words, illustrated in Table 2. Interestingly, Hand C seldom goes wrong in monosyllables which Orm would spell with a double consonant (11%), but they make mistakes 42% of the time when the relevant segment occurs in a multisyllabic word. Examples such as < underr > (Orm. unnderr), with correct -derr but incorrect un-, suggest that in their attempt at applying Orm’s spelling system it might have been difficult for Hand C to be mindful at once of all segments of multisyllabic forms. Monosyllables ending in < VC_iC_i > were presumably easier to spell correctly because they are shorter and because most are frequently-used function words or verb forms whose spellings would have been easier to pick up due to constant use.

Table 2 Hand C’s record in heeding Orm’s model regarding syllable- and word-final single and double consonants in mono- and multisyllabic words

Hand C’s orthography exemplifies precisely the messiness and confusion that Orm probably wished to remedy, and although they tried to follow Orm’s model, time and again they reverted to their own conventional orthography.¹⁶ Crucially, Hand C’s orthographic system did not involve a consistent association of double final consonants and short preceding vowels. For instance, there is no reason to suspect that they pronounced third person singular present indicative forms with a short preceding vowel in < tredeþþ > and with a long preceding vowel in < endeþ > and < lerneþ > (cf. also < wiþ > on 43r and < wiþþ > on 62r).

Unexpected Mistakes

In light of the observed tendency concerning the direction of mistakes (i.e., the preference for erroneous CC > C rather than C > CC) and the lesser degree of difficulty Hand C had with monosyllables, the appearance of a double consonant in < herr > ‘here’ (< OE hēr) and < fótt > ‘foot’ (< OE fōt) is all the more surprising. However, there is reason to believe that < herr > is a true misspelling, while < fótt > may be deliberate.

Neither hēr nor fōt had in OE been regularly spelled with a double consonant, and neither form with double consonant was frequent in early Middle English. A search for < herr > in the Dictionary of Old English Corpus (DOEC) returned zero results; the only attestation in the sense of the adverb ‘here’ in the Corpus of Middle English Prose and Verse (CME) is the object of the present discussion. There are no attestations in the Helsinki Corpus, and the Oxford English Dictionary does not list < herr > as an attested form (OED Online, 2022). By contrast, a double-consonant form < fott > is attested elsewhere, though not exactly with great frequency. Searching the DOEC for this variant form of Orm’s főt (< OE fōt ‘foot’) yields two occurrences: one from Owun’s tenth-century Northumbrian¹⁷ gloss of the Rushworth Gospels, and one from the Life of St Margaret (ed. Clayton & Magennis, 1994, pp. 112–138) in London, British Library, Cotton MS Tiberius A. iii (s. xi med., cf. Ker, 1957, no. 186).¹⁸ Schlemilch (1914: 62–63), who collects late OE (c. 1000–1150) forms with unexpectedly-doubled final consonants following a long etymological vowel, likewise finds attestations of < fott > , but not of < herr > ; indeed, he finds none with final < rr > . Unlike < herr > , < fott > seems to be a spelling which a contemporary scribe might have used deliberately.¹⁹

The manuscript context, too, provides support for the idea that < herr > is a mistake but < fótt > – at least in Hand C’s mind–is not. While < fótt > (H4978a²⁰/43r) occurs within a verse rewritten by Hand C on the basis of its deleted equivalent in Orm’s hand found on the inner margin of the same folio, < herr > (H19616/117v) is part of a section of text added by Hand C in the lower margin without the possibility of recourse to Orm’s model on the same folio.²¹ The word < herr > is immediately preceded on Orm’s final line of the column by < broþerr > , the final < err > -segment of which might have caused Hand C, who was struggling to imitate Orm’s orthography, to misspell her by analogy; the three half-lines preceding Hand C’s addition furthermore contain a total of five syllable-final < Vrr > -sequences, and no syllable-final < Vr > . A different situation presents itself with respect to < fótt > . Hand C had Orm’s version of H4978 right before them on 43r, and they still committed several blunders in their rewriting of said version (cf. Johannesson, 2007: 117). Yet the presence of the acute accent on < fótt > betrays their awareness of Orm’s version: nowhere else does Hand C place accents, and it is unlikely that they should independently decide to endow this particular word with a single acute, considering also that single acutes do not occur on (or on the vowels before) < tt > -segments elsewhere in the Ormulum. Rather, it seems Hand C understood that there was something special about főt, which did not however induce them to copy Orm’s word letter by letter (and diacritic by diacritic).

Single acutes conventionally and in Orm’s own usage indicate a long vowel; a single acute could also mark stress, but it would be surprising if Hand C had marked stress only in this instance, in a monosyllable no less, where there is no doubt about the (non-existing) internal stress hierarchy, and in the final position of a first half-line, which is always a stress position in the Ormulum and thus poses no difficulty at all. Since Hand C’s orthography lacked a general association of double final consonants with a short preceding vowel (see above), this single acute which probably indicates a long vowel need not have contradicted–in Hand C’s mind–the final < tt > . It seems likely, then, that Hand C’s spelling < fótt > for Orm’s főt is deliberate.

An Orthographic Habit in Orm’s House

If it is correct that < fótt > was a sensible spelling in Hand C’s mind and that their single acute indicates a long vowel, it follows that Hand C’s < tt > sequence cannot at the same time signal a short preceding vowel–the single acute would stand in obvious contradiction to the linear orthography. Hence, whatever the doubling of the consonant graph in < fótt > expresses, it cannot relate to preceding vowel length.

Orm, who probably conducted an “extensive preparatory study of English writings” (Anderson & Britton, 1999, p. 306) in order to arrive at his reformed spelling system, would have been aware that single final consonants could be preceded in OE by long or short vowels, whereas a double final consonant would normally signal a short preceding vowel–he was merely the first person of whom we are aware to apply the rule systematically.²² In other words, vowels preceding single final consonants could be viewed as underspecified in OE with respect to preceding vowel length, but before double final consonants a vowel was (usually) short. Knowing this, Orm might have taken issue with a tendency in his fellow canons’ writings–as exemplified by Hand C’s orthography–of doubling final < t > after a long vowel in violation of this rule he perceived and would go on to generalize in the Ormulum.²³ He might have worried that potential copyists would fail to observe the contrast between single and double final < t > , a danger of which Orm may have been acutely aware in light of the predominance of monosyllables ending in < tt > rather than < t > in the Ormulum, with a ratio of roughly 16:1.²⁴

Orm might have believed that special marking of < t > via double accents would remedy this problem. The lack of multi-accenting on most of Orm’s < tt > sequences could then be explained by the direction of the tendency found in Hand C: they substitute < tt > for < t > , not generally the other way round, and in the latter case it would have been irrelevant for what reason < t > was doubled. In summary, the presumed inclination of his immediate audience to double final < t > in words with long preceding vowels could have prompted Orm to prophylactically counteract via double accents. And although Hand C did not follow Orm’s spelling convention when they wrote < fótt > , they apparently made a compromise that saw them both retain final < tt > and acknowledge the length of the preceding vowel.

A French Connection?

The hypothesis laid out above is admittedly speculative, based as it is on the presence of a single form, but it is a curious coincidence at the very least that Hand C so markedly deviated from Orm’s model found on the same folio. It appears useful now to consider possible origins of this spelling habit which may have been present in Orm’s house.

Schlemilch (1914, pp. 62–63), examining the orthography of late Old English texts (c. 1000–1150), proposes that such double spellings are evidence of consonant lengthening. He may be right, but the double consonants may also express something else entirely, or they may be completely meaningless–we simply do not know, and on this point I shall remain neutral pending further study. Of interest are the forms listed for t (cf. Schlemilch, 1914, p. 62),²⁵ because every single one has a multi-accented analogue in the Ormulum:

(5)	a.	ū̆tt (independently and as a prefix; Orm. űt < OE ūt ‘out’)
	b.	fē̆tt (Orm. fe̋t < OE fēt ‘feet’)
	c.	fō̆tt (Orm. főt < OE fōt ‘foot’; cf. Hand C < fótt >)
	d.	mō̆tt (Orm. mőt < OE mōt ‘be_allowed_to.prs.3sg’)
	e.	forlē̆tt (Orm. forrle̋t < OE forlēt ‘abandon.pst.sg’)

It becomes evident from Schlemilch’s lists that (i) the phenomenon is significantly more common for < t > than for any other consonant, and that (ii) it is more common from 1100 to 1150 than from 1000 to 1100.²⁶ He finds < tt > -spellings in four manuscripts, three of which are dated to the twelfth century: the Textus Roffensis (c. 1122–1124), the Codex Wintoniensis (c. 1130–1150), and CCCC MS 383 (s. xiiⁱⁿ).²⁷ The one exception is a single attestation of forlē̆tt which occurs in the OE interlinear gloss of the Rule of St. Benedict in Cotton MS Tiberius A. iii (s. xi med.); the MS may predate the Conquest, but not by much.

A few generations post-Conquest, when three of the four manuscripts mentioned above were produced, the scriptoria of England would have been staffed with scribes who had in their training been exposed to French writings to a much higher degree than their pre-Conquest predecessors. But certain Old English texts retained some relevance and continued to be copied, such as the law codes from the Textus Roffensis and CCCC 383, or–for obvious reasons–the cartulary of Winchester Cathedral Priory in the Codex Wintoniensis. If a new class of unexpected spellings surfaces in such texts it stands to reason that the French language itself or the habit of writing French manifests itself in this change.²⁸ Worley (2003, pp. 23–24) points out that Orm’s spelling system would have been useful in counteracting changes in pronunciation that affected the French language in Orm’s day, and it seems one must indeed have recourse to the infamous ‘Anglo-Norman scribe’ in order to account for the problematic spelling tendency that might have occasioned Orm’s counter-measure, even if doing so follows in the frowned-upon tradition of blaming any strange variation or inconsistency in the spellings of Middle English texts on ‘Anglo-Normans’ (cf. Milroy, 1992, pp. 193–196; Clark, 1992 dedicates a whole chapter to the ‘Myth of the Anglo-Norman scribe’).

Many of the French consonants (such as t) underwent phonological changes just prior to or during the twelfth century. In Old French, “t and d were [in certain positions] pronounced ‘th’ as in ‘thin’ and ‘then’ respectively, […] in which case they had disappeared in pronunciation and usually in spelling by the early twelfth century” (Einhorn, 1974, p. 5). The positions to which Einhorn refers are final (after a vowel) and intervocalic, and following Laborderie (1994, pp. 63, 68) the development happened in parallel for both consonants and positions. Other final consonants were lost, too, but -t and -d preceded -k, -f, -s, -l, -r, -n, and -m by half a century or more: while the loss of final t and d was complete by c. 1100, the others followed suit only from the second half of the twelfth century and well into the thirteenth (Fouché, 1952, p. 663). Anglo-French tends to preserve certain archaic spelling features longer than Continental French, among them intervocalic and (apparently less frequently) final t and d, but the corresponding dental fricatives were in most cases either not pronounced or we have no clear evidence that they were (Short, 2013, §24.1–4).²⁹ It is not necessary to suggest that Orm considered his target audience of canons or other preachers so slow-witted that they would require constant reminding of the pronunciation of t (but not d, strangely) in English. Orthography, however, is a different question.

When it became necessary for French scribes–or scribes trained on French–to write English, the French association of < t > and < d > with the dental fricative impacted their spellings. According to Schlemilch (1914, pp. 56–58), final < t > had occasionally been used in late Old English to designate the fricative normally represented by < þ > or < ð > , but this use was much strengthened after the Conquest; at the same time, though much less frequently, < ð > was also substituted for < t > in initial and final position, < d > alternates frequently with < ð > in every position, and < t > and < d > in final position alternate frequently as well. Interestingly, it appears from Schlemilch’s description that < t > was only used to represent a fricative finally, while < d > , < ð > and < þ > occurred in various positions (cf. also Schlemilch, 1914, p. 53). This is the rule also in the Peterborough Chronicle’s Final Continuation (for the years 1132–1154). In it one may find both < t > and < d > in final position, with < th > , < þ > and < ð > being preferred initially and medially, even though the Final Continuator was “English-speaking and knew the native symbols þ and ð” (Clark, 1970, p. lxiv). It can be inferred that Orm himself–even if the consistency of his reformed spelling mostly obscures the fact–was not immune to this influence, spelling the name of the biblical figure ‘Abihu’ once < abyud > (H539/15r) and once < Abyuþþ > (H480/13r); the spelling in < d > may have come from either Latin or French, but the form ending in < þ > shows that Orm would have likely pronounced either one with a final dental fricative.³⁰

The ambiguity of < t > in twelfth-century English texts could have prompted Orm to draw special attention to this letter, and the apparent limitation of < t > as a spelling for the dental fricative to final positions–following Clark (1970, p. lxiv) due to “spirantal pronunciation being commonest there in Norman”–could explain Orm’s lack of interest in and accenting of the letter elsewhere. As noted above, however, final < d > does not usually receive multi-accenting in the Ormulum (unlike final < t >), which requires explanation given that < d > could also be pronounced as a fricative in Anglo-French and was used by the Peterborough Chronicle’s Final Continuator and others to represent that sound.

Perhaps the origin of the divide between Orm’s treatment of < t > and that of < d > lies in an etymological factor which resulted in the retention of a plosive pronunciation in Old French for some instances of final -t but not for -d. Importantly, final -t resulting from the simplification of the Latin geminate consonant spelled -tt- (and the loss of a final vowel) continued to be pronounced (Fouché, 1952, p. 661; e.g., Latin cattus > OFr chat ‘cat’).³¹ Other consonants descending from Latin geminates were retained also, but for these the ambiguity of the consonant graph was not an issue as early on because their analogues not descending from geminates were lost much later than previously fricativized -t and -d. Crucially, descent from the Latin geminate -dd- does not seem to have saved any final -d from effacement in Old French (cf. Fouché, 1952, p. 661). It may be interesting to note, moreover, that the set of consonant graphs which Schlemilch (1914, pp. 62–63) finds unexpectedly doubled after OE long vowels is almost identical with the subset of Latin geminates whose simplified, word-final descendants evaded effacement in Old French. Table 3 illustrates the pattern. The first column lists all consonants found at least a hundred times word-finally in the Ormulum (singly, after a vowel; plus < g > and < b > to match the Latin geminates), examples being provided in the second column; in column three, the existence of Latin geminates using the respective letter is indicated; column four shows which of these Latin geminates’ simplified descendants seem to have avoided a loss of pronunciation in Old French (cf. Fouché, 1952, p. 661); and column five reproduces words with the corresponding final consonant unexpectedly doubled after a long OE vowel as listed by Schlemilch (1914, pp. 62–63).

Table 3 Overlap of the set of consonant graphs unexpectedly doubled as found by Schlemilch (1914) with simplified Latin geminates in final position in Old French

It is thus the case that, when readers of Continental French or Anglo-French manuscripts in Orm’s day saw a final < t > , they could pronounce (i) a plosive, (ii) a fricative, or (iii) nothing at all; when they saw final < d > , their range of options was limited such that they could only pronounce either (i) a fricative, or (ii) nothing at all.³² But there was a second graphie besides < t > that could represent /t/, namely < tt > . It furthermore lacked the ambiguity of < t > , for although the Latin geminates mentioned above had been simplified, the double spellings familiar from Latin (in this case < tt >) were often retained (cf. Scragg, 1974, p. 50). According to Pope (1952, §1218), this conservatism–not specifically for /tt/ but with respect to all kinds of Latin spelling models–was cultivated both in Continental and Anglo-French. When Anglo-Norman scribes saw < tt > – surviving from the spelling of the Latin intervocalic geminate -tt-–they would therefore associate it exclusively with the plosive /t/.

To summarize, it seems there existed a difference in the range of pronunciations available for final < t > on the one hand, which could be (i) a plosive, (ii) a fricative, or (iii) nothing at all, and final < tt > on the other, which could only be pronounced as a plosive. It does not seem unreasonable to suppose that the French association of < tt > with /t/ alone could have induced Anglo-Norman scribes to write < tt > for final < t > in English (whether habitually or consciously as a sort of note-to-self that an English word had a /t/-sound rather than a /θ/), just like they sometimes substituted < t > for final < þ > based on French convention. The pattern from Table 3 above would fit nicely with the idea of a spelling habit arising out of the intermixing of French and English traditions–including the hypothesis regarding the prevalence of /t/ in Schlemilch’s list as a consequence of the corresponding sound’s early effacement in French.

Whether this speculative account of the introduction of double consonant spellings into English words with an etymologically long preceding vowel is correct or not, it is noteworthy that “[t]he association [of double consonants] with preceding short vowels is in origin English, beginning in the late Old English period when long vowels were shortened before a combination of two following consonants” (Scragg, 1974, p. 50). By contrast, it was, according to Pope (1952, §1170), not until the later period of Anglo-French that “quantitative differences appear to have been gradually established […], mainly on the lines of the English quantitative differences, i.e., long vowel in open syllables, short vowel in blocked ones […]”. This usage postdates the Ormulum, and a twelfth-century French-trained scribe need not therefore have considered the presence of final < tt > contradictory to a long pronunciation of a preceding vowel–something which has also been posited above for the Ormulum’s Hand C, and which has been identified as, if followed by a copyist, ruinous to the integrity of Orm’s orthographic system.

A very brief remark of Skeat’s (1901, p. 473) likewise suggests that there was something going on with final /t/ as French and English traditions cross-fertilized:

The E. [English] final t sounded differently. I fancy it sounded to them [i.e., Anglo-Norman scribes] stronger, with a sort of emphatic final splutter. Hence we find leth for let (he let). This th is by no means our E. th, but a t with an explosive sound after it, like lett’. We even find thown for town.

Schlemilch (1914, p. 56) does not find many instances of th substituting for t in late Old English, but Skeat’s assertion is not far-fetched at all, for the English t was alveolar, and the Anglo-French t dental (cf. Pope, 1952, §§1115, 1113).³³ Note also that Skeat’s example is leth (OE lēt), which appears as double-accented le̋t in the Ormulum; a similar segment–not with the same meaning, of course–also appears in Orm’s inconsistently-spelled < onndle̋tt > / < onndlæ̋t > ‘face’ (< OE/Anglian ondwleata). Might Orm have fallen back, just like Hand C when they spelled < fótt > rather than főt, into a French-influenced habit from before the conception of his reformed spelling when he put to parchment the double-consonant variant–the < tt > perhaps expressing just like < th > “a sort of emphatic final splutter”–despite the long preceding vowel he had in mind? Assuming that English and Anglo-French /t/ did indeed sound different, it is not impossible that scribes perceiving this would have sometimes felt the need to render English final /t/ as < tt > after a long vowel. Such practice could have reinforced, or could have been reinforced by, an independent habit of doubling final < t > (originating in the attempt to conserve Latin spellings of former geminates) perhaps transferred from French to English in the spelling of Anglo-Norman scribes. And finally, it may be interesting to note that the atypically multi-accented segments in the Ormulum do not have in common a long vowel, a < t > , or indeed a final consonant, but a letter which in twelfth-century English texts could represent an alveolar plosive (< d > , < t > , < tt > , < th >) or a dental fricative (< d > , < t > , < th >), which also points to the spelling of /t/ rather than vowel length as the issue that was truly the thorn in Orm’s side.³⁴

Conclusion

Evidence has been put forward in the present study to suppose that Orm’s double accents do not immediately and redundantly indicate vowel length. It disagrees on this point with earlier explanations of Orm’s double accenting by, among others, Trautmann (1896), Luick (1914/1964), Sisam (1933, 1953/1962), and Markus (1989). The Ormulum’s Hand C misspells Ormian főt ‘foot’ as < fótt > , from which it has been inferred that < tt > did not necessarily indicate a short preceding vowel to them. Orm might have wished by means of his double acutes to counteract a spelling habit–the doubling of final t after an etymologically long vowel–present in his immediate audience of bilingual or purely francophone canons which stood in obvious contradiction to the basic rule of his reformed orthography requiring vowels preceding double final consonants to be short. Unexpected doublings of this kind seem to be a phenomenon more frequent after the Conquest than before it, occurring more often for t than for any other consonant, and some evidence has been adduced on the basis of which I speculated that such a spelling habit could have entered English via French-trained scribes at a time when French and English scribal traditions cross-fertilized. We do not know whether the difference between final < t > and < tt > after a long vowel was somehow reflected in the pronunciation of the scribes using the latter spelling, but even an essentially meaningless spelling habit could have seemed sufficiently problematic in Orm’s mind to prompt a systematic counterreaction. Further study of twelfth-century interactions between the English and (Anglo-)French phonologies and orthographies remains very much a necessity, for example in regard to the curious near-identity between the sets of simplified Latin geminates in Old French and the consonants Schlemilch (1914) finds doubled finally in late Old English.