skip to content


Early Middle English for today

full menu

simplification of verb paradigms in eME

some complexity falls from the OE verbal system

The changes in late OE and eME verb paradigms in the East Midlands, can be summarised as follows:

Note: unless specified otherwise, eME denotes the normalised early East Midland used in this site.

core vs optional grammar

The changes to verb paradigms which were complete by 1100 pass to both the core grammar and the optional grammar of eME. Changes which were underway in early Middle English pass to the core grammar but not to the optional grammar. For a detailed discussion, see Normalisation 3: grammar.

The core grammar, which has fewer anomalies and exceptions, is aimed at the beginner. This is the grammar that appears in Audrey and the attercop, book 1 of Englisc buten taeres (the first 800 words). The optional grammar, which inherits more of OE's anomalies and exceptions, appears in normalised versions of OE and ME texts, and in the weblog of this site.


assimilation of present tense forms

In OE, most strong verbs (except class 1) as well as class 2 weak verbs, had a dual stem in the present tense. Assimilation of the two different stem vowels begins in the OE period.

In eME, in both cases, a single stem emerges:

Note that the assimilation of the present tense forms works in different directions for strong and weak verbs.

assimilation in strong contracted verbs

In the core grammar, assimilation of present tense forms in strong contracted verbs, entails the following:

In the optional grammar, assimilation of present tense forms in strong contracted verbs, entails the following:

For more detail on strong contracted verbs, see below.

unaccented vowels

The levelling of unaccented vowels to schwa /@`/, observed in other parts of speech, also affects verbs, so that:

Unstressed <e> in verbs

Elision of the unstressed <e> in the 3rd person singular present and both 3rd person preterite forms, is optional:

back to top

extending the pattern in verbs

East Midland sources are fairly conclusive on the loss of <i> in the stem of class 2 weak verbs. Thus we have Orm lufenn and Ch love(n). We also see assimilation in verbs from the OE's class 3 weak verbs - PC2 saein and Orm sey%y%enn sey%y%t`, where OE had sec`g`an saeg`th. We see assimilation as far back as OE itself in class 2 weak verbs with <g`> between the stem and ending, e.g. sme_an as well as standard sme_ag`an15. That trend continues in eME with PC1 freon (< fre_og`an).

However, the evidence for the concurrent loss of gemination in class 1 verbs is thin. There aren't any examples of verbs like fremen in Orm16 or PC2 or Ch. Nevertheless, TH (Homilies in Cambridge, Trinity College), a text that dates from 1200, has fremen in the sense 'perform, accomplish'.

We know then that the assimilation we see in class 2 and class 3 verbs, also existed to some degree, in class 1 verbs, around the year 1200. On that basis, I've applied principle 4 of the five basic principles. That is - I've extended a levelling pattern evident in the three classes of weak verbs, to all weak verbs, in order to simplify the verb paradigms of eME.

Note that because there is no explicit support for the dropping of gemination in either Orm or PC2, gemination in class 1 weak verbs remains in the optional grammar.

back to top

simplification of present tense forms in strong verbs

In OE, the vowel of the second- and third-person singular forms of the present tense of strong verbs often differed from that in the infinitive. Most OE grammars list strong verb paradigms which show only forms with i-mutation: lu_can s2 ly_ct`, helpan s3 hilpt`, stelan s4 stilt`, sprecan s5 sprict`, flo_wan s7 fle_wt`.

However, i-mutation of the second- and third-person singular of strong verbs was by no means universal. Boswell cites the following examples, (in addition to mutated forms): D`u__ monegum helpst...He__ helpet` t`earfan parcet...d`u__ bru__cest... he bru__cet`... Gif frigman fre__um stelt`... sprecet`... flo__west... flo__wet`. The extent of the confusion is perhaps most starkly illustrated by sle__an: slehst slieht` sliht` slaeht` are all recorded in Boswell.

This complexity disappears altogether by the beginning of the ME period in the East Midlands, but can still be seen here and there in the early 13th century in the West Midlands and Southern dialects. Thus, while AW still has kimet`, Orm has cumesst cumet`t`. The assimilation of 2nd and 3rd person present tense forms is widespread in pre-1250 texts: 1225 helpet`, 1250 helpest and Orm beresst, beret`t`, bruket`t`, dellfet`t`, fallet`t`, hatet`t`, speket`t` to cite just a few more examples.

Note also Orm biddet`t`, bitet`t`, ridet`t`, singet`t` where OE often had syncope, especially after a dental consonant: bitt, bi_tt, ri_tt, singt`.

In fact the simpler paradigm in which the vowels of all present tense forms match their respective infinitives, is so firmly entrenched in early ME sources that there is no need to call on principle 4. Since i-mutation of 2nd and 3rd person forms is missing in all relevant strong verbs in both Orm and PC2, this grammatical feature doesn't pass to the optional grammar.

back to top

OE contracted verbs ("h-less")

OE contracted verbs are those strong verbs which once had an <h> in the infinitive form, in Primitive Old English, but which had lost this by the 8th century, the era of our earliest surviving records of the written language. That <h> survived in the 2nd and 3rd person singular preterite form, right up to Chaucer's time. These "h-less" infinitives can be found in 5 different strong classes:

Side-by-side comparison of strong verbs - standard vs contracted

These contracted verbs present a couple of anomalies or complexities which are not found in the standard paradigms of their respective classes, as the following comparisons attest. Note that the i-mutation of the 2nd and 3rd person present tense forms in both standard and contracted types, is not dealt with in the observations below:

class 1

ri_dan s1
inf. ri_dan
pr sg1 ri_de
pr sg2 ri_dest (ri_tst)
pr sg3 ri_det` (ri_tt)
pr pl ri_dat`
pt sg3 ra_d
pt pl ridon
p ptc riden
wre_on s1 (contracted)
inf. wre_on
pr sg1 wre_o
pr sg2 wri_hst
pr sg3 wri_ht`
pr pl wre_ot`
pt sg3 wra_h
pt pl wrigon
p ptc wrigen

In the class 1 strong contracted verbs, we see two anomalies:

  1. <h> is missing from the end of the stem in the infinitive, present 1st person singular and present plural forms;
  2. the stem vowel in the infinitive, present 1st person singular and present plural forms, is <e_o> rather than <i_>;

classes 2 & 5

The class 2 and class 5 strong contracted verbs can be treated together:

be_odan s2
inf. be_odan
pr sg1 be_ode
pr sg2 be_odest (by_tst)
pr sg3 be_odet` (by_t)
pr pl be_odat`
pt sg3 be_ad
pt pl budon
p ptc boden
te_on s2 (contracted)
inf. te_on
pr sg1 te_o
pr sg2 te_ohst (ti_ehst)
pr sg3 te_oht` (tieht`)
pr pl te_ot`
pt sg3 te_ah
pt pl tugon
p ptc togen
sprecan s5
inf. sprecan
pr sg1 sprece
pr sg2 sprecest (spricst)
pr sg3 sprecet` (sprict`)
pr pl sprecat`
pt sg3 spraec
pt pl sprae_con
p ptc sprecen
se_on s5 (contracted)
inf. se_on
pr sg1 se_o
pr sg2 sehst (siehst)
pr sg3 seht` (sieht`)
pr pl se_ot`
pt sg3 seah
pt pl sae_gon (sa_won)
p ptc segen (sewen)

In the class 2 and class 5 strong contracted verbs, we see a single shared anomaly:

  1. <h> is missing from the end of the stem in the infinitive, present 1st person singular and present plural forms;

class 6

bacan s6
inf. bacan
pr sg1 bace
pr sg2 bacest (bae_cst)
pr sg3 bacet` (bae_cet`)
pr pl bacat`
pt sg3 bo_c
pt pl bo_con
p ptc bacen
sle_an s6 (contracted)
inf. sle_an
pr sg1 sle_a
pr sg2 slehst (sliehst)
pr sg3 slaeht` (slieht`)
pr pl sle_at`
pt sg3 slo_h
pt pl slo_gon
p ptc slagen (slaeg`en)

In the class 6 strong contracted verbs, we see two anomalies:

  1. <h> is missing from the end of the stem in the infinitive, present 1st person singular and present plural forms;
  2. the stem vowel in the infinitive, present 1st person singular and present plural forms, is <e_a> rather than <a>;

class 7

gangan s7
inf. gangan
pr sg1 gange
pr sg2 gangest
pr sg3 ganget`
pr pl gangat`
pt sg3 ge_ng
pt pl ge_ngon
p ptc gangen
fo_n s7 (contracted)
inf. fo_n
pr sg1 fo_
pr sg2 fe_hst
pr sg3 fe_ht`
pr pl fo_t`
pt sg3 fe_ng
pt pl fe_ngon
p ptc fangen

In the class 7 strong contracted verbs, we see three anomalies:

  1. <h> is missing from the end of the stem in the infinitive, present 1st person singular and present plural forms;
  2. the stem vowel in the infinitive, present 1st person singular and present plural forms, is <o_> rather than <a>;
  3. the <n> of the preterite and past participles stems, is missing in the infinitive and present tense stems;

First thoughts arising from these comparisons

As stated above and at basic principles, my goal is to simplify verb paradigms within a related group of eME verbs, wherever a levelling pattern is evident in pre-1250 MED sources. As we have just seen, there are anomalies in the paradigms of contracted strong verbs, which involve both the stem vowel and the absence of a final stem consonant.

The question is - to what extent is levelling of these anomalies in the paradigms of contracted verbs, supported in ME?

As it turns out, there is considerable evidence of levelling of contracted verbs in pre-1250 sources. Already in OE (perhaps more so in the Anglian dialects), <h> is dropped at times from 2nd and 3rd person present tense forms. That is increasingly evident in ME. On the other hand, <g> is occasionally inserted at the end of the stem in infinitives and in 1st person and plural present tense forms. This begins in OE and continues in ME where <y`> (or <h>) often replaces the <g> spelling. We'll examine this in more detail below.

<h>, <g> and <y`>

Bear in mind that <h> /h/ doesn't occur between vowels in OE or ME 5. So when I talk below about extending (or restoring) <h> throughout the present tense paradigm of an eME reflex of an OE contracted verb, that extension can only be via a related sound. As we'll see in the examples below, <h> is generally restored as /G``/ with a <g> or <y`> spelling 6. Note that this levelling could also be seen as an extension of <g> (or <ng>) from the preterite and past participle forms to the present tense.

In the following selection of entries drawn from the MED, you'll see a range of verb forms for each of the sixteen strong contracted verbs, with a focus on pre-1250 sources. Among them is a fair smattering of <g> (or <h>), in both infinitive and present tense stems.

The OE strong verbs affected are:

Selection of MED entries for each OE contracted strong verb

le_on s1 ligen p ptc
to lend, grant
li_h imp sg la_h pt sg3 ligen p ptc;
The MED has lenen v w inf from lae_nan v w inf, but not *le_n/*li_en/*li_gen v s1 inf.
OS far-li_han, OHG li_han, Goth leihwan
si_gan (se_on/si_on) s1 sigen p ptc
I to sink, descend; II to strain, filter, ooze, trickle
sa_h sigon pt sigen p ptc
MED si_en inf; 1200 syht`, 1381 seth pr sg3; AW siy`et`, 1225 seiy`eð pr pl; 1225 seh pt
Note - both Clark Hall 1 and Bosworth & Toller 4 list si_gan with the two definitions above, and se_on with the latter definition only, while Wright 3 identifies se_on (strain) as one of the contracted class 1 strong verbs, separately from si_gan (sink).
te_on (ti_on) s1 tygen/togen p ptc
to accuse
ty_hst ti_ht` te_ot` pr te_o sbj pr sg te_ah tugon pt tuge sbj pt sg tygen/togen p ptc;
MED te_n (1250 tegen) inf; 1450 tyxste pr sg2, 1300 tiy`th pr sg3, 1150 teo sbj pr sg;
Note - in OE this verb seems to have largely shifted to the conjugation of te_on s2 (from pOE *te_ohan)
OS af-ti_han, OHG zi_han, Goth teihan
t`e_on s1 t`igen/t`ogen p ptc
to thrive, flourish, prosper
t`i_hd` pr sg3 t`e_ot`/t`i_gat` pr pl t`e_o sbj pr sg t`e_onde pr ptc t`a_h/t`e_ah t`igon/t`ugon pt t`igen/t`ogen p ptc;
MED the_n (1250 d`en) inf; 1175 d`ihd` pr sg3; Ch then inf theech (thee ich) pr sg1; TH 1200 t`ied` pr sg3; 1150 t`eah pt s3, 1250 d`ogen pt pl/p ptc
OLG/OS thi_han.
Note - this verb has a foot in at least two camps. Forms consistent with classes 1, 2 and class 3 are recorded. For more detail, see footnote below. The normalised eME in this site uses only class 1 forms with this verb.
wre_on (wri_ga) s1 wrigen p ptc
to wrap, cover
wre_o wri_hst wri_gd` pr sg wre_od` pr pl wre_ah (a_) wrigon pt;
MED wri_en (Lmn wriy`en, AW wrihen, 1250 wrihe) inf; AW wrihe pr sg1; AW wrihd` pr sg3, SO 1330 writ`, Ayb 1340 wriy`t`; AW wrihed`, Ayb 1340 wret` pr pl; 1175 wriy`en pt pl; Lmn iwriy`en, 1300 wriy`e p ptc
OHG int-ri_han
Note - the MED cites an alt OE infinitive form wri_gan, which Clark Hall 6 doesn't list; however Bosworth & Toller 7 has an entry for wri_ga verb (to cover).
fle_on (fli_on) s2 flogen p ptc
to flee
fli_onde pr ptc fle_o fli_hst/fly_hst fli_ht`/fly_ht` pr sg fle_ot`/fli_ot`/fly_t` pr pl fle_ah fluge flugon pt
MED fle_n (Orm fle(o)n, Lmn fleon, AW fleo, 1325 fle, 1400 fleiy`en) inf; Owl flyhst/fliy`ste pr sg2; Orm fle(o)t`, Lmn flet`/flict`, Owl fliy`t, 1250 flet` pr sg3; Lmn fle(o)nde, 1175/1200 fleonde pr ptc; Lmn fliy`/fle(o), AW flih imp sg; 1250 fledde, 1300 fle(d)de, 1325 fled, Ch 1375 fledde pt sg3; Lmn floy`en pt pl; Lmn 1300 ifloy`e, 1325 fled, Ch 1395 fled p ptc
OS fliohan, OHG fliuhan, Goth t`liuhan
te_on s2 togen p ptc
to draw, tow, tug; draw near, approach
te_o ty_hst tieht` (ti_ht`/te_ht`) te_ot` pr te_oh imp sg te_onde pr ptc te_ah tugon pt
MED te_n (Lmn teon, AW teon, 1400 tegh) inf; Lmn tid`, Owl tiht`, 1200 ted`, 1300 ted` pr sg3; AW teod`, 1250 tgen, SO 1330 tet` pr pl; Owl teo pr sbj sg; Lmn taeh pt sg3; 1150 tugen, Lmn tuy`en, AW tuhen pt pl; 1250 togen p ptc
OS tiohan, OHG ziohan, Goth tiuhan
(g`e)fe_on s5 fegen (faegen) p ptc
to rejoice
feogad` pr pl feah fae_gon pt
not found in MED
ple_on s5 *plegen p ptc
to risk, expose to danger
pleah pt
not found in MED
se_on s5 sewen/segen p ptc
to see
se_o siehst sieht` (si_t`) se_od` (si_ot`) pr seh (i/io) imp sg seah sa_won/sae_gon pt
MED se_n (Orm sen/seon, Owl ison) inf; Orm seo, Owl iso, 1335 se pr sg1; Orm sest (eo), Lmn sehste, Owl sichst, 1275 sixst, 1300 sext/sixt/sicst pr sg2; PC1 sed` (eo), Orm set` (eo), 1175 sihd`, 1225 sicd`, Owl sut`, 1250 sikt`, Ch seth, 1396 sey`t pr sg3; Orm seo, Lmn sey`e, Owl so sbj pr sg; Orm sen, 1250 sen, 1300 seth, Ayb 1340 zyet` pr pl; Lmn isaeh pt sg3; 1200 segen, 1250 sey`en, 1275 seghen, 1300 seiy`en pt pl; 1335 seiy`en p ptc
Note - the MED lists seiy`en and sey`e as infinitive forms, seghe as pr sg1, and say`e, saghe, sauy`en as pt pl, at the top of its article on se_n.
OS sehan, OHG sehan, Goth saihwan
fle_an s6 flagen p ptc
to flay
*flo_h *flo_gon pt
MED flen (Lmn flan/flean, 1275 flo, 1300 flon, 1300 fleye) inf; AW flea sbj pr sg, 1333 flagy`st pr sg2; 1390 fleth pr sg3; 1300 Hav flowe, 1350 flogh pt sg3; Lmn iulay`ene, 1300 yflawe, 1340 Ayb vlay`e, Ch flayn p ptc
ON fla_, MDu vla`en vla`eghen, ProtoG *flahana
le_an s6 *lagen p ptc
to blame, reproach
lyht` le_at` pr lo_h lo_gon pt
not found in MED, but Wikt. cites lough pt sg3
ON la_, OS/OHG lahan, Goth laian
sle_an s6 slaeg`en/slagen p ptc
to slay, strike
slehst (sle_s) pr sg2 slieht`/slae_ht`/sle_t` (i/y) pr sg3 sleah imp sg sle_ande pr ptc slo_h slo_gon pt
Orm slaen/slan inf slast slaet`/slat` pr sloh sloy%h`enn pt slay%enn p ptc, Lmn slaen inf slaest slaed` pr sloh sloy`en pt, AW to sleanne inf dat slead`/slad` pr sg3 sloh slohen pt slein/slayn p ptc, SO slo inf slouy` pt sg3, Ch sleen inf sleest sleeth sleen pr slough/slow slowen pt slayn/slawen p ptc, MED sle_n (1300 sla, 1325 slan, 1340 Ayb slay`e, 1390 slawe) inf; 1325 slecst, 1325 slas, 1340 Ayb slast, 1372 slest pr sg2; 1225 sleihd`, 1303 slekt`, 1340 Ayb sslay`t/slay`t`, 1325/1340 slas, 1380 slet` pr sg3; 1225 LH slage, 1300 slext`, 1357 sla pr pl; 1350 slay` sbj sg pr?; 1340 Ayb slay`et` imp pl; PC2 sloghen pt pl; 1200 TH slaine, 1250 slagen, 1307 slaye, 1300 slan, 1325 slain, 1340 slayne, 1350 sleyn/slaw, 1393 Gower slawe p ptc
Notes - the spelling <y`> in Ayenbite of Inwyt (1340 Ayb) represents /G``/, e.g. foly`en, while <y> represents /j/, e.g. yeve, daye; the MED lists A sla_ inf, lOE slagen p ptc and slay`en inf at the top of its article on sle_n, the latter perhaps based on Ayb slay`e.
ON sla_, OS/OHG/Goth slahan
t`we_an s6 g`et`wegen 7 p ptc
to wash, cleanse
t`we_a d`wehst/t`wyhst d`weht` pr t`weh/d`uah t`we_ad` imp t`we_a t`we_an sbj pr t`wo_h t`wo_gon/d`wo_gan pt
MED thwe_n inf; 1175 d`weahd` pr sg3; 1200 t`weh imp sg; 1175 it`wey`en p ptc
ON t`va_, OS thwahan, OHG dwahan, Goth t`wahan
fo_n s7 fangen/fongen p ptc
to seize, grasp
fo_ fe_hst fe_(h)th/fo_hth fo_th pr fe_ng fe_nge fe_ngon pt fo_h fo_t` imp fo_ fo_n sbj pr fe_nge fe_ngen sbj pt
MED fongen/fo_n (Orm fon, fanngenn, Lmn fon, AW underuongen, Owl/1225/1250/1300/1303/Ch fonge, 1275 fong, PH3 foangen, 1300 fange) inf; Orm fot`, 1150 fot`, 1200 TH fod` fohd`, AW underued`, Owl ifot`, 1333 fanget`, 1393 Gower fonget` pr sg3; Owl fo pr pl sbj; Orm unnderrfanngenn, 1390 fongen, 1393 Gower fonge pr pl; PC2/AW underfeng, PC1/1175/Lmn feng pt sg3; Orm onnfengenn, Lmn on-fengen pt pl; PC1 gefangen p ptc
ON fa__/fanga, OS fahan/fangan, OHG fa__han, Goth fahan
ho_n s7 hangen p ptc
to hang
ho_t` pr pl ho_h imp sg he_ng he_ngon pt
MED hongen (Lmn hon, AW hon/hongin, Owl hongi, Ch hangen/honge) inf; Ch hange pr sg1; Lmn hangest pr sg2, AW hongi sbj pr sg2; Owl hot`, AW honged`, 1250/1307/1325/SO 1330/Ayb 1340 honget`, 1303/Ch hanget` pr sg3; Ch hangen pr pl; Orm/Lmn/Ch heng pt sg3; PC2 hengen pt pl; Ch hangen p ptc
Note - there is no record in the OE corpus of present tense singular forms for ho_n. OE also had hangian w intr (to hang, be hanged).
ON hanga, OHG hahan, Goth hahan

Levelling of paradigms for strong contracted verbs in ME

In the following discussion, <g/y`> represents either /G``/ or /J``/, the former following a back vowel and the latter following a front vowel. It is a fricative, and not the palatal consonant /j/ which is also sometimes represented by <y`> in ME, and by <y> in ModE.

As mentioned above, there are several examples of the dropping of <h> in 2nd and 3rd person present tense forms in early ME. The dropping of <h> will be touched on but I'm more interested initially in the retention of <h> and more importantly - the extension of <g/y`> to the infinitive as well as first person and plural present tense forms. The reason - the result is in line with the standard paradigm of the respective verb class, in which the final stem consonant of the present tense forms reflects the final stem consonant of the preterite and past participle forms.

From the pre-1250 sources in the verbs listed above, there are:

  • five examples of a <g/y`> (or <h>) extended to the infinitive: OE si_gan, wri_ga, 1250 tegen, AW wrihen, Lmn wriy`en
  • seven cases of <g/y`> (or <h>) extended to plural present tense: OE t`i_gat`, feogad`, AW siy`et`, 1225 seiy`ed`, 1250 tgen, 1225 LH slage, AW wrihet` pr pl
  • one case of an <h> extended to a 1st person present: AW wrihe
  • six cases of <ng> extended to an infinitive, or any present tense form (class 7): Orm fanngenn, AW underuongen, hongi, Lmn hangest, AW honged`, 1250 honget`

back to top

Normalising strong contracted verbs in ME

In the early stages of learning a language, the key is to offer the simplest route. This is where principles 3 & 4 come into play. These allow us to generalise a levelling trend that is underway. The upshot is that while the present tense of OE strong contracted verbs have a dual stem, their reflexes in the core grammar of eME, have a single stem.

The approach is as follows:

  1. we start with a verb paradigm that exhibits stem alternation, e.g. wre_on wre_o wri_ht` iwrigen;
  2. in a pre-1250 text we find a form which removes the stem alternation for either the inf or pr sg1 or pr pl; e.g. Lmn wriy`en inf;
  3. via principle 3 the stem levelling of one form (e.g. inf) is extended to the remaining forms (e.g. pr sg1 and pr pl) within the paradigm; so wri__gen inf is extended to wri__ge pr sg1 and wri__gen pr pl;
  4. courtesy of principle 4, the paradigm levelling of the majority of strong contracted verbs can be extended to the remaining verbs in that group; e.g. the wri__gen wri__ge wri__get` iwrigen pattern is extended to li__gen li__ge li__get` iligen;

Thus we have:

pre-1250 removal of stem alternation:
OE si_gan, wri_ga, 1250 tegen (< *ti_han), AW wrihen, Lmn wriy`en inf; OE t`i_gat`, feogad`, AW siy`et`, 1225 seiy`ed`, 1250 tgen (< *te_ohan), 1225 LH slage, AW wrihet` pr pl; AW wrihe pr sg1; Orm fanngenn, AW underuongen, hongi, Lmn hangest, AW honged`, 1250 honget` class 7 inf or pr
via principle 3 - extend removal of stem alternation throughout paradigm:
wri__gen8, ti__gen, t`i__gen s1; te__gen s2; fegen s5; *slagen s6; fangen, hangen s7
via principle 4 - extend pattern of simplified paradigm to remaining verbs in group:
li__gen s1; plegen s5; lagen, t`wagen s6
special cases - contracted forms that have survived to ModE:
fle__n s2, se__n s5, slaen, flaen s6


  1. with few exceptions, all forms of the special cases - fle__n, se__n, slaen, flaen qualify individually as additional eME forms:
  2. Orm flen flet`, Lmn fleon flet`, 1250 fledde pt sg3 presage forms we are familiar with in ModE flee flees fled;
  3. h-less infinitive forms are well supported in pre-1250 sources for se__n, slaen and flaen: Orm sen slaen, Lmn slaen flean, AW seon sleanne, Owl ison;
  4. similarly, h-less 2nd and 3rd person singular forms are well supported in pre-1250 sources for se__n and slaen: Orm sest set` slaet`, Trinity Homilies 1200 sest, Lmn slaest slaed`, AW slead`
  5. it seems most likely that ModE slay, flay continue the Midland h-less forms of Orm slaen, Lmn slaen, AW sleanne and Ch sleen rather than the Southern (South-Western) and Kentish forms Lambeth Homily 1225 slage, Ayenbite 1340 slay`e /G``/, which would have resulted in the ModE form slawe; the addition of a palatal glide or dipthongisation from /E``/ to /ei/ giving sleye sleyth sleyn i.a., doesn't appear in the infinitive and present tense forms until around 1390; hence the core eME forms slaen, flaen;
  6. Orm slay%enn , AW/Ch slayn p ptc establish islay%en as an additional eME form;

which leads us to:

Core grammar - conjugation of OE contracted verbs in eME

Stem alternation is removed in all related verb paradigms.

The final consonant of the preterite plural stem - <g>, is extended to the present tense forms. The only exceptions are the four special cases, which are familar to ModE speakers:

  • s1 eME wri__gen inf (wrap, cover), wri__ge wri__gest wri__get` wri__gen pr, wra__h wrigen iwrigen
    • similarly: li__gen (lend, grant), si__gen (sink; strain), ti__gen (accuse), t`i__gen (thrive)
  • s2 eME te__gen inf (draw, tow), te__ge te__gest te__get` te__gen pr, taeh tugen itogen
  • s5 eME fegen inf (rejoice), fege fegest feget` fegen pr, fah fagen ifegen
    • similarly: plegen (risk, endanger)
  • s6 eME t`wagen inf (cleanse), t`wage t`wagest t`waget` t`wagen pr, t`wo__h t`wo__gen it`wagen/t`way%en
    • similarly: lagen (blame)
  • s7 eME fangen inf (seize, grasp), fange fangest fanget` fangen pr, fe__ng fe__ngen ifangen
    • similarly: hangen

Four contracted verbs which have survived to ModE, depart from the patterns above in eME. They retain the contraction of their respective OE forms and extend the dropping of <h> at the end of the stem throughout the present tense. Note also that fle__n has weak past tense forms:

  • s2 eME fle__n inf (flee), fle__ fle__st fle__t` fle__n pr, fledde fledden ifled (fled)
  • s5 eME se__n inf (see), se__ se__st se__t` se__n pr, sah sagen isegen (saw seen)
  • s6 eME slaen inf (slay), slae slaest slaet` slaen pr, slo__h slo__gen islagen/slay%en (slew slain)
    • similarly: flaen (flay)

Optional grammar - conjugation of OE contracted verbs in eME:

The present tense forms in classes 2 and 5 are based on standard OE forms. Class 1 forms are based on the additional eME form wri__en via ModE obs. wry and 1250 wrien as well as the expected Anglian form arising from primitive OE <i_h> (more on that below). Principle 2a (Orm + Lmn/AW/Owl) supports the dropping of <h> and the assimilation of the stem vowel in the 2nd and 3rd person forms of classes 6 and 7 - Orm slaet` fot`, Lmn slaest slaed`, AW slead`, Owl ifot`.

In the list below, the infinitive appears initially for each verb. Otherwise, the only forms given are those which differ from the core grammar forms above:

  • s1 eME wri__en inf, wri__e wri__hst wri__ht` wri__en pr
    • similarly: li__en, ti__en, t`i__en
  • s2 eME te__n inf, te__ te__hst te__ht`9 te__n pr
    • similarly: fle__n, with s2 past forms: flaeh flugen pt iflogen
  • s5 eME se__n inf, sehst seht` pr, saegen pt pl
    • similarly: fe__n, ple__n
  • s6 eME t`waen inf, t`wae t`waest t`waet` t`waen pr
    • similarly: laen
  • s6 eME sla__n (slaen) inf, sla__ sla__st sla__t` sla__n pr
  • s7 eME fo__n inf, fo__ fo__st fo__t` fo__n pr
    • [similarly: ho__n]


  1. si_en s1 (OE WS se_on) is not included here as si_gan covers both meanings - sink/descend and strain/filter in Clark Hall;
  2. re class 1 strong contracted verbs - OE WS <e_o> appeared as A <i_>; see Wright 14 § 127: "i_ was broken to i_o before h and ht in WS. But already at an early period the i_o mostly became e_o (= Anglian i_)". Examples cited by Wright include: WS we_oh = A wi_h (idol), le_oh (lend) imp sg = A li_h and with loss of medial h after breaking had taken place: WS fe_ol = A fi_l (file), li_on/le_on (to lend), si_on/se_on (to strain), t`i_on/t`e_on (to thrive), wri_on/wre_on (to cover); cp Ch wryen wrien;
  3. re class 1, 6 & 7 strong contracted verbs - alternation of the stem vowel isn't found in either the core or optional grammar of eME; see the related discussion above on i-umlaut in Simplification of present tense forms in strong verbs;
  4. Orm sla slat`, AW slad` and SO slo support additional eME forms sla__n sla__t`, and by extension through principle 3 - sla__ sla__st; the sla__n forms are derived from ON sla__;
  5. present tense forms with stem hang- are recommended for ho__n; see discussion at optional grammar

back to top


Gemination occurs in both strong and weak verbs in OE. For both strong and weak verbs it entails the doubling of the stem-final consonant in the infinitive and present participle as well as the first person and plural forms of the present tense, the subjunctive forms and the plural imperative form.

The tables and verb listings below are taken from Wright 11:

class I weak verbs with gemination

inf. fremman
pr sg1 fremme
pr sg2 fremest
pr sg3 fremet`
pr pl fremmat`
pr ptc fremmende
sbj sg fremme
sbj pl fremmen
imp sg freme
imp pl fremmat`
pt sg3 fremede
pt pl fremedon
p ptc g`efremed
inf. settan
pr sg1 sette
pr sg2 setst
pr sg3 set(t)
pr pl settat`
pr ptc settende
sbj sg sette
sbj pl setten
imp sg sete
imp pl settat`
pt sg3 sette
pt pl setton
p ptc g`eseted/set(t)

The difference between these two class I verb types is that syncope (dropping of <e>) and assimilation (of <d> to <t>) occurs in the preterite and past participle of verbs with stem-final <tt> in the infinitive. Thus *setede becomes sette.

Like fremman are conjugated aswebban to kill, clynnan to sound, cnyssan to knock, dynnan to make a noise, gremman to anger or provoke, hlynnan to roar, hrissan to shake, scet`t`an (also strong) to injure, swet`t`an to swathe, temman to tame, trymman to strengthen, t`ennan to stretch, t`ic`g`an† (in poetry also strong pt t`eah, t`ah) to receive, wec`g`an† to agitate, wennan to accustom, wret`t`an to support.

Like settan are conjugated atreddan to search out, cnyttan to bind or knit, hreddan to rescue or save, hwettan to whet or incite, lettan to hinder, spryttan to sprout, lec`g`an† to lay.

† Note that in the case of <g`> /j/, the doubling is represented by <c`g`> /dZ`/.

Note also the similarity with class 2 weak verbs: the gemination is absent in the 2nd and 3rd person singular present tense forms and in the preterite forms.

OE variants and ME developments

The following list provides variants to the geminate verbs above, listed under the corresponding OE verb. The variants comprise entries from two OE dictionaries - Clark Hall 12 and Boswell-Toller 13 as well as pre-1250 extracts from the MED. The MED examples focus on East Midland entries up to and including Chaucer. I've highlighed entries which display gemination either missing or applied where not expected. Note that in the list below, CH refers to Clark Hall 12 and BT to Boswell-Toller 13.

fremman, fremian, framian 14
avail, benefit, advance; perform, accomplish
fremian, framian; note also - fremu n f, freme adj frem~sum ~lic`
AW fream~ien ~et` pr sg3, StK frem~ien ~ede pt sg3, PH3 freme, HA frem~ed` ~ad` pr, BH frem~aet` ~iaet` pr ~ode pt sg3, Lmn fremmen (v), HA to y`efremmenne, BH fremmaen fremede pt sg3 ifremed p ptc, TH fremen, MED fremmen, fremen, framen.
all three OE variants < Proto-Germanic *framjaną; cognate with OS fremmian, OFris frema, fremma
pre-1250 missing gemination: fremian, framian, fremen fremien freamien fremad` fremiaet`
put to sleep, lull; † kill
g`eswefian put to sleep, lull, appease; note also - swefan v w intr 'sleep, slumber, rest'; † 'sleep in death', swefn (swefen) n nt 'dream'
VH 1225 sweueð pr sg3 y`eswefede, 1225 swefede, Lmn sweuede pt sg3 isweued, MED sweven put to sleep; become calm, isweven put to sleep, stupefy
pre-1250 missing gemination: g`eswefian
† lull, soothe, set at rest; put to death, destroy
Ch asweved p ptc, MED asweved
(make a) sound
clynnan CH, clynan BT
MED - no entries
pre-1250 missing gemination: clynan
cnyssan CH & BT
cnyssan/cnysan cnysset`/cnyset` pr sg3 cnyssed p ptc cnyssedan/cnysedon/cnysdan pt pl BT
BH 1175 cnysed`, MED knissen
pre-1250 missing gemination: cnysan
pre-1250 added gemination: cnysset`, cnyssed, cnyssedan
dynian CH & BT
make a noise
[dynnan = dynian] CH, dynet`; note also - dyne n m
Lmn dunien dunede pt sg3, AW dunede pt sg3, Owl adunest pr sg2, MED dinen
pre-1250 missing gemination: dynian, dunien
gremian CH & BT
irritate, provoke
[gremman = gremian] CH, gremiad`
PC1 to gremienne, Lmn graemende, AW gremien gremed` pr sg3 gremede pt sg3, MED gremen anger, provoke, offend
pre-1250 missing gemination: gremian, gremienne, graemende, gremien
sound, make a noise, shout
[hlynian = hlynnan] CH, hlynnet` pr sg3 BT
MED - no entries; homonym in ME with linnen 'cease, desist'
pre-1250 missing gemination: hlynian
pre-1250 added gemination: hlynnet`
hrisian CH & BT
shake, move; be shaken, clatter
hryse imp sg hrisiendis pr ptc? hrisedon; hrissan? Wright/MED/Wikt.
Lmn rusien riseden, AW risede, Ch rese inf, MED resen tremble, shudder, quake
> ModE rese obs. 'shake; quake; tremble'
pre-1250 missing gemination: hrisian, rusien
sc`ed`d`an w/s6?
scathe, injure, hurt, crush, oppress, disturb
sc`et`t`e sc`et`et`/sc`et`t`at` pr sg3 sc`et`ede pt sg3 BT; sc`(e)o_d pt sg3 Wright; sceat`ian injure, spoil, steal, [sceat`an = sc`et`t`an] CH; sc`ead`~a n m ~e f
Orm skat`esst pr sg2 skat`edd p ptc, Cl 1380 scat`ed pt pl, HA 1200 scet`t`es imp sg, VH 1125 scad`ied` pr pl, MED scathen (k), shathien, scet`t`en
pre-1250 missing gemination: sceat`ian, sceat`an, scad`ied`
pre-1250 added gemination: sc`et`t`at`
swad`ian CH
wrap, swathe
swet`ian swet`ede BT, beswed`ian CH & BT, swet`t`an? Wright
VH 1125 swad`ede pt sg3, MED swathen, swat`elen
pre-1250 missing gemination: swat`ian, swet`ian
temian CH & BT
temige temet` temat` temede temedon temma BT
1333 y-tamed p ptc, Gow tamed, MED tamen
pre-1250 missing gemination: temian temat`
trymian CH
strengthen, fortify, comfort
trymman/trimian BT, trymman trymmend trymmung trymnes trymd` CH
1250 trimen trimede, VH 1150 trymede, BH 1150 trymmen, MED trimmen exhort, encourange, strengthen
pre-1250 missing gemination: trymian trimian, trimen
stretch, extend; prostrate
t`ennan/t`enian d`enige BT, t`ennan (stretch) t`ynnian (thin, dilute) CH
1400 thinne, 1450 thyn pr sg1, MED thinnen < OE t`ennen; extend
pre-1250 missing gemination: t`enian
receive, consume; beg
t`icgan CH & BT, t`ah t`eah and t`igde t`igede pt sg t`ae_gon t`e_gon and t`i_gdon t`igedon pt pl t`egen and t`iged p ptc BT
VH 1150 to t`ycgene inf t`ycge sbj sg t`ycgden pt pl (/j/?), 1175 to t`icgene, 1200 HA t`iy%ean/t`icy%an inf t`iy%est d`iy%t` pr t`icy%e (/j/?) t`ey%e imp sg, 1225 LH t`igge sbj sg, 1300 Hav t`igge inf, MED thiggen
pre-1250 missing gemination: t`iy%ean
pre-1250 added gemination: t`ycgden pt pl, t`icy%e imp sg
wec`g`an, wagian 17
move, agitate, shake, wag, move backwards and forwards
wecgan CH & BT, wagian CH & BT, weget` wecgat` wecge wagiat` wagat` BT;
1425 weged, 1440 wedge (wegge), MED weggen to wedge into place, tighten; 1187 Wagetail, 1219 Waggestaf, 1226 Waggespere, AW waggid`, Hav 1300 wagge, 1330 wagge, Ayb 1340 wagget`, Ch waggen, MED waggen tr/intr move back and forth or up and down, shake, flutter, swing
both OE variants < Proto-Germanic *wagjaną
pre-1250 missing gemination: wagian wagende NB - no 'i'
pre-1250 added gemination: waggid`, Waggestaf, Waggespere
wenian CH & BT
accustom, habituate, inure, train; entertain, treat; break off, wean from
[wennan = wenian] CH, wenian wenede BT (accustom oneself, be accustomed)
AW iwenet, p ptc, MED wenen wean; entertain
pre-1250 missing gemination: wenian
wred`ian CH & BT
support, sustain, uphold
wred`ian CH, wret`ian wret`et` wret`iat` wret`edon BT (prop, stay, support, sustain)
MED - no entries
pre-1250 missing gemination: wret`ian

OE sec`g`an w lec`g`an lic`g`an byc`g`an fric`g`an t`ic`g`an wec`g`an

In OE, the paradigms of these verbs display stem alternation. While the infinitive, 1st person singular and 3rd person plural forms all contain <c`g`>, the 2nd and 3rd person singular forms have <g`> in its place, e.g sec`g`an sec`g`e saeg`st saeg`t` sec`g`at`. In the early Middle English period, levelling is evident in these paradigms.

Infinitive forms with diphthongs in /ai/,/ei/ (or contraction to /i:/) rather than /edZ`/,/idZ`/ are attested in PC2 - saein lien. Neither Orm or PC2 has a form similar to ley%y%enn or leien 3. On the contrary, Orm has leggen. However pre-1250 examples of /ei/ rather than /edZ`/ exist: 1175 We leged` 10 and 1200 Ne mai no mann leiy`en. Thus there is a clear pattern here, which we can extend courtesy of principles 3 & 4. We can be reasonably confident that if say%en, liy%en and ley%en were spoken by East Midlanders around the year 1200, then the same applies to other verbs with a similar stem alternation, e.g. biy%en 11 friy%en 12 t`iy%en.

Thus, alongside the forms derived from OE - secy%en lecy%en licy%en bicy%en fricy%en t`icy%en which enter the optional eME grammar, we have the levelled forms - say%en ley%en liy%en biy%en friy%en t`iy%en, in the core eME grammar. These give us simpler verb paradigms without stem alternation, e.g. say%en say%e say%st say%t` say%en, say%de in the core grammar, while the optional grammar retains the stem alternation of OE - secy%en secy%e say%st say%t` secy%en, say%de.

The outlier is the pair wecy%en/waggen17 which have separate senses in the core grammar: to wedge and to wag (shake, agitate, move backward and forward) respectively. In the optional grammar both forms share the second sense - to wag (shake, agitate, move backward and forward). The ModE sense 'to wedge' is rendered in the optional grammar by dón/maken wecy%. Note also that wecy%en is the lone verb in this group which extends <cy%> rather than <y%> throughout the paradigm in the core grammar. This reflects not only the ModE form to wedge but also the likelihood that wecy%en is a derivative of wecy% (< OE wec`g` n m) rather than a reflex of OE wec`g`an v w.

back to top

syncope in preterite forms

Sycope of the unstressed vowel in the preterite suffix was common in OE weak verbs, particularly in certain environments. The exception was the class II group of weak verbs.

This distinction fell away in eME. For example, OE lufode becomes lufde in Lmn.

back to top

double consonant becomes single in preterite

This affects ll, mm, nn, pp, rr.

Example: OE fyllan fylde g`efylled

Similarly: OE cennan, c`ierran, clyppan, cyssan, fiellan, mierran, spyllan, stillan, wemman

With one exception - the variant preterite spilt (alongside spilled), this change to the root in the preterite is not reflected in ModE.

This degemination persists to 1400. For example, Ch has kiste as well as the less common variant kissed.

On the other hand, retention of the double consonant begins early. HA 1200 has y%efullede pt sg3. And alongside custe we find cussede in Lmn and cussed in AW.

Given that the retention of the double cononsant in preterite forms is evident in pre-1250 texts, this rule - double consonant + <ed> = single consonant + <t> in preterite forms, has been shifted to the optional grammar.

back to top

ferien, herien

We saw above that in ME, the infinitive and 1st person and plural present tense forms of class 2 weak verbs, are assimilated to the preterite and 2nd and 3rd person singular present tense forms. OE verbs with infinitives ending in ~ian, lose the <i> in ME. Their ME reflex has an infinitive ending in ~en, and all forms in the paradigm follow suit.

ferien (to transport, carry) and herien (to praise) are the two exceptions to this rule. In their case, assimilation takes place in the reverse direction: the preterite and 2nd and 3rd person singular present tense forms are assimilated to the infinitive form. Since their infinitive ends in ~ien, all forms - preterite included, have the <i> in their stem.

In other words, where OE has herian inf herie herast herat` heriat` pr herode pt sg3 g`eherod p ptc heriende pr ptc, eME has herien inf herie heriest heriet` herien pr heriede pt sg3 iheried p ptc heriende pr ptc.

There is a reason for the exeptional treatment of ferien and herien. These two verbs have <i> in their stems in eME because they have retained the <i> throughout the paradigm in either Ch or ModE or both. Ch has herie inf heriest pr sg2 heryed pt sg3 iheryed p ptc, while ModE has herry (obs.) and ferry ferries ferried 13. The MED lists them as ferien and herien

back to top


  1. with two exceptions - ferien and herien (see last section above)
  2. also - OE byc`g`an > Ch byen
  3. Ch has leye as well as leggen
  4. Clark Hall 2 indicates that (g`e)t`e_on belongs to both class 1 and class 3. The Wiktionary page on t`e_on indicates class 3 above its conjugation, and gives the etymology as: "From an earlier (unattested) form *þīhan, from Proto-Germanic *þinhaną, from Proto-Indo-European *tenk-. Cognate with Old Saxon thīhan, Old Dutch thīon (Dutch gedijen), Old High German dīhan (German gedeihen), Gothic 𐌸𐌴𐌹𐌷𐌰𐌽 (þeihan)." The present tense forms are shown as: t`ig`e t`i_st t`i_ht` t`i_g t`i_gat`. The past tense forms are shown as: t`a_h t`ugon g`et`ogen/t`ig`en. The Wiktionary page on Proto-Germanic *þinhaną, tells us that the verb is strong class 3 and that "After -inh- had become -į̄h-, this verb was shifted to the first strong class by analogy in most daughter languages, but not in Old English. Old English also shows a secondary formation with leveling of Verner's alternation, probably back-formed to the lexicalized original past participle ġeþungen (grown, thrived)." Apparently this Wiktionary information is taken from Kroonen, Guus (2013) Etymological Dictionary of Proto-Germanic (Leiden Indo-European Etymological Dictionary Series; 11), Leiden, Boston: Brill. The Cambridge Old English Reader 5 lists g`et`e_on in its glossary with the indication "I/II", i.e. strong classes 1 and 2. Back to t`e_on »
  5. <h> is found between vowels in AW but represents /G``/
  6. conceivably levelling could have been via <hh> (/x/), since the <h> of the 2nd and 3rd person forms precedes a consonant, and in that context is pronounced /x/ rather than /h/; cp hliehhan, pohhe
  7. Boswell-Toller lists four different p ptc forms for t`we_an - t`wagen, t`waeg`en, t`wegen, t`wogen near the top of the article; however, of these only g`et`wegen can be found in the examples themselves;
  8. wri__gen s1 (wrap, cover) has a near homonym in wriy%en w intr (turn, twist, bend); note that while Clark Hall has wri_gian Boswell-Toller lists the latter as wrigian and it may have been wrig`ian (> ModE wry); similar near homonyms are liy%en s5 (lie)/adj (flaming), t`iy%en (t`icy%en) w (thrive), tey%en w (tie), fe__gen (fe__n) w (hate, persecute)
  9. pr sg2/3 forms with <h> prevent some confusion, for example - te__ht` pr sg3 (draws) vs te__t` n pl (teeth)
  10. PM leged` probably represents /lei@`T``/ since in PM /dZ`/ is generally spelt <gg>, e.g. buggen seggen, and intervocalic <y`> usually stands for /G``/, e.g. fuy`eles, ay`en
  11. bi(y`)en is not found in the MED prior to 1300; early instances include: 1300 beye, 1325 bij, 1330 bien, 1372 byy`e, 1378 bie, 1390 byen, Ch bye (all inf)
  12. Orm has friy%y%en and Hav (1300) has frie, but both in the sence of taunt, find fault rather than ask; see MED frien which gives the source as ON fryja (question, taunt, find fault)
  13. cp ModE bury buries buried; birg`an w1 > Orm birry%enn, but is treated like a class 2 weak verb at times: Lmn burede, ibured, bibured and 1300 burede, i-bured, alongside PC2 bebyried, Lmn biburiyede and AW iburiet; Ch has buryen inf burieth pr sg3 yburyed p ptc
  14. three verbs or one? Clark Hallref 1 lists separately fremman 'to further, advance, support; frame, make, do, accomplish, perfect, perpetrate, commit, afford', fremian 'to avail, benefit, do good' and framian 'to avail, benefit'; the sense of 'perform' is evident in OE fremman (Bosworth-Toller ref 4); the MED distinguishes fremen 'aid, advance, benefit', framen 'benefit, profit' and fremmen 'perform, accomplish'; TH 1200 fremen (listed under MED fremmen) has the sense 'perform, accomplish'; in Shakespeare frame has the sense 'execute, perform': All have sworn him an oath that they should frame his will on earth; some users may prefer to separate framen avail, benefit from fremen perform, accomplish;
  15. similar OE class II weak verbs are fe_og`an (fe_on) fre_og`an (fre_on) twe_og`an (twe_on) t`re_ag`an (t`re_an);
  16. if Orm had an instance of fremmenn, the <mm> would indicate a doubled consonant (rather than a short vowel) since Orm only used the doubling of a consonant after a short vowel in a closed syllable which was final or preceded a syllable starting with a consonant, e.g. godd, follgenn; cp Orm witenn pt pl in which the <i> is short, but doesn't trigger a doubling of <t> because the following syllable starts with a vowel;
  17. for wec`g`an CH gives 'move, agitate, drive hither and thither; be moved' while BT has 'wag, move, shake'; for wagian CH gives 'move, shake, swing; totter', while BT has 'wag, wave, shake, move backwards and forwards; totter; be loose'; Wikt. has conflicting etymologies: ModE wedge secure using a wedge; drive into a narrow gap < OE wec`g`an move, agitate, drive hither and thither and OE wec`g` wedge, mass of metal, lump; note the significant difference in meaning between the two OE sources; it seems that OE wec`g`an and wagian were variants descended from one PG verb *wagjaną, and initially covered a related set of meanings but gradually diverged and specialised to produce ME wedge and wag; that differentiation is already evident in ME;


  1. Clark Hall J.R., A Concise Anglo-Saxon Dictionary, 4th edn (Cambridge, 1960)
  2. Clark Hall ibid.
  3. Wright, Joseph & Wright, Elizabeth Mary, Old English Grammar, London : H. Milford, Oxford University Press, 1914, pp. 255
  4. Bosworth, Joseph. "se_on." In An Anglo-Saxon Dictionary Online, edited by Thomas Northcote Toller, Christ Sean, and Ondřej Tichy. Prague: Faculty of Arts, Charles University, 2014.
  5. Richard Marsden, Cambridge University Press, 2004
  6. Clark Hall op. cit. p. 364
  7. Bosworth, op. cit. "wre_on"
  8. Bosworth, ibid. "t`we_an"
  9. Bosworth, ibid. "wrigian"
  10. Clark Hall op. cit. p. 364
  11. Wright op. cit. pp. 267-9
  12. Clark Hall op. cit.
  13. Bosworth, op. cit.
  14. Wright op. cit. p. 63