• Dutch
  • Frisian
  • Saterfrisian
  • Afrikaans
Show all
Strong and other irregular verbs
quickinfo

Next to the regular or weak verbs, Frisian has quite a number of irregular verbs. These could be divided into four groups, although the division is not without intersections.

Most irregular verbs can be characterized as strong. Their main feature is ablaut. As a result of this historical event, the vowel of the stem varies: the past tense and/or past participle may show a different vowel than the infinitive. This is different from the regular weak verbs where the vowel of the stem remains constant. An example of a strong verb is the verb slute to close, which has /y/ in the stem. Its past tense is sleat, with /I.ə/; the past participle is sletten, with /ɛ/. The vowel alternation is sufficient to indicate the past tense, and hence we see that the strong verbs lack the suffixes of the weak verbs which do mark them. Instead of the suffixes for the past participle (-d/-t for class I, -e for class II), we see the suffix -en as a second main feature of the strong verbs. An example is the participle sletten closed, which can be analyzed as slet-en.

Frisian also has weak verbs that are nevertheless irregular in that they show stem allomorphy. Irregularity emerged after historical phonological processes that affected a part of the paradigm. An example is meitsje to make. In Old Frisian this was makia, in those days fully regular (as for instance Dutch maken is still a regular weak verb). The /i/ (or maybe later /j/) of the ending caused palatalization, both of the preceding consonant /k/ and the stem vowel. The result is an infinitive form meitsje, and a present tense paradigm which runs as follows: (ik) meitsje I make - (do) makkest you make - (hy) makket he makes and (wy/jimme/hja) meitsje (we/you/they) make. The past tense makke and the past participle makke are regular, at least in the light of the Old Frisian forms. If, on the other hand, we were to take the infinitive meitsje and its putative derived stem meits- as point of departure, these forms would be irregular. Meitsje is a verb belonging to class II of the weak verbs. Another subgroup in this class displays an alternation /j~g/, for instance in the opposition jei vs. jage in the present versus past tense of the first person singular of the verb jeie to hunt. Furthermore, we find several weak verbs of class I that partially underwent a historical phonological shortening, which succeeded in a paradigm which is comparatively irregular. An example is the verb briede to roast, with its first person double ik bried I roast.PRES vs. ik brette I roast.PRET.

A third group of irregular verbs consists of auxiliary verbs. As in other languages, the most irregular verb is wêze to be. Another important category are the modal verbs, former preterite-present verbs. A characteristic feature is that they do not show a suffix -t in the third person present tense. So English 'he can' is Frisian hy kin, and not *hy kint.

Monosyllabic verbs constitute a fourth irregular category. Their infinitive lacks an ending -e or -je. Moreover, these verbs often show different stems. The verb sjen to see, for instance, has a present tense on the basis of the stem sjoch-.

A minor irregular subgroup consists of the verbs sizze to say and lizze to lay. Their irregularity is also caused by historical palatalization processes. Finally, a minor group of verbs shows irregularity in the present tense in that the vowel of the form of the second and third person of the singular is shortened (and often lowered). An example is the verb bite to bite, which has /i/, where its second and third person show /I/. This alternation is obsolete nowadays, but it still occurs in the dialect of the island of Schiermonnikoog.

Next to their inherent irregularity, strong and irregular verbs may display other peculiarities as well. For instance, auxiliary and monosyllabic verbs may have shortened forms, dealt with below in the section on apocope. An example is the present tense stem doch- of the verb dwaan to do, which may drop the final fricative, becoming do-. We also see developments with respect to the suffixal endings. One is the result of a historical rule of metathesis, which affected the past plural suffix -en. The order -ne can still be found after certain stems ending in a vowel, as described in past tense stems ending in a vowel. The plural past tense of the verb hawwe to have, for instance, is hie-ne, and not something like *hie-en. We also see attempts to make the endings more transparent. One can be located in the west of the language area, where strong participles that have absorbed the ending -en in the stem are augmented with a suffix -d, probably transferred from the weak verbs of class I. Thus the past participle dien of the verb dwaan to do surfaces as diend in these dialects. A relatively new development is the reinforcement of the ending -st of the second person singular of the past tense. This may turn into -est by insertion of a schwa. For instance, we get the form do tochtest you thought, which was do tochtst before.

As might be expected, there is pressure to reduce irregularity. Levelling occurs at various places. In particular, deviating forms in the present and past tense are vulnerable.

Finally, this topic also contains lists of strong and irregular verbs.

readmore
[+]Strong verbs

As the other Germanic languages, Frisian has a considerable number of verbs, more than 100, that originate in the Germanic ablaut system. They lack the dental suffixes that build the preterite and the past participle. Instead, the stem of one or both of these categories displays a different vowel. An example is slute to close. Its preterite is sleat, and sletten is the form of the past participle. A total of three different vowels is rather rare, however. Thus the vowel of the infinitive may return in one of the other forms, for instance ride-ried-riden to ride, with a deviating preterite. Or, as is often the case, preterite and past participle display the same vowel, as in skinke-skonk-skonken to give, to pour.

A second feature of strong verbs is the ending of the past participle, which is -en, as may be seen in the examples above. In various cases, this ending is less conspicuous in that it is integrated in the stem syllable. If this is the case, the schwa of the suffix deleted. Examples are bedoarn spoiled, bleaun remained, n tied and sean cooked. The ending -en may even be the only indication that a verb is strong, at least historically. This is the case with the participle laden loaded, from the infinitive lade to load, which has a weak past tense lade loaded.

A few strong verbs also show variation in their consonantism, due to historical phonological processes. The verb sjitte-skeat-sketten to shoot has a deviating onset in the infinitive and present tense, as a result of palatalization before a high front vowel; in Old Frisian the infinitive was skiata. Palatalization also yields a different onset in the verbs jitte to pour, ferjitte to forget and jilde to cost, be valid (cf. English yield). Compare, for instance, the preterite geat and the past participle getten with the infinitival form jitte. Apart from the prefix, ferjitte to forget displays the same forms. The verb jilde has the non-palatalized form gou as a preterite. The verb sjonge-song-songen to sing has a similar onset as sjitte to shoot, but here as a result of some form of Old Frisian breaking (see Boutkan (1998). Something comparable has affected the first member of the triple fjochtsje-focht-fochten to fight.

Furthermore, a couple of verbs with an infinitive stem ending in /k/ show palatalization and assibilation in the past participle (which must have occurred in a period in which the following ending had a front vowel; cf. older spellings like -in). This concerns the participles brutsen (from brekke to break), dutsen (from dekke to cover), lutsen (from lûke to pull), rutsen (from rekke to stretch), sprutsen (from sprekke to speak), stutsen (from stekke to stick), strutsen (from strike to iron) and trutsen (from trekke to pull). Around the year 1600 the form baetsen still occurred, from bakke to bake, a verb that later turned to the weak class in its entirety. In the first time after this palatalization, the original final /k/ remained intact in the preterite, but later, especially in the south, the preterite forms of all of the members of this cluster took over the rhyme of the stem form of the past participle. For instance, for many speakers in the north the verb brekke to break still has the pattern brekke-briek-brutsen, but in the south this is generally replaced by brekke-bruts-brutsen.

Palatalization effects can also be observed at the end of the stem of the verbs sizze to say and lizze to lay, this time affecting the voiced velar consonant. It resulted, for instance, in a fricative /z/ in the infinitive (via intermediate /dz/), and in /j/ in some other forms. This resulted, among others, in the triple sizze say.INF to say, sei said.1SG.PRET /saj/ and sein said.PP /sajn/. Synchronically, no phonological relation between these various sounds exists anymore. As these verbs are originally weak, and moreover show other peculiarities, they will be dealt with in the verbs sizze en lizze in a separate section below.

A few verbs show a peculiar behaviour. On the one hand, they show vowel alternation, which points to their being strong. This idea might be reinforced by the fact that the past tense and past participle even show an extra consonant, i.e. /χ/. An example is bringe to bring, which has the form brocht brought both as past tense and past participle. Here is an overview of the verbs that have the same pattern:

Table 1
Infinitive 1/3SG.PRET Participle Translation
bringe brocht brocht bring
keapje kocht kocht buy
sykje socht socht seek
tinke tocht tocht think
Note that the past participle ends in a dental stop, which rather reminds one of what is usual in the weak verbs. The present tense of these verbs is regular. This means that bringe and tinke are conjugated along the lines of the weak class I, and keapje and sykje along class II. The modal auxiliary meie may shows a comparable pattern, with its past forms mocht may.1/3SG.PRET and mocht may.PP. There are differences, however. Firstly, although the orthography is the same, the vowel of mocht is one grade higher, i.e. pronounced with [o] instead of [ɔ]. Furthermore, the past participle has an alternative form meien. And as a modal auxiliary, the third person singular of the present tense behaves differently since it lacks the inflectional ending -t, as may be seen in auxiliary verbs below.

Next to the unexpected final cluster /ɣt/ we also find a deviating final cluster /st/ in wist, the preterite of witte to know, and in doarst, one of the forms of the past tense and the past participle of doare dare. For the latter, see also the section on auxiliary verbs.

Scattered remnants of Verner's law may be found in variation in the coda of the verbs frieze to freeze and ferlieze to lose. They have as preterite the forms frear and ferlear, in which /z/ turned to /r/.

An alphabetical list of the Frisian strong verbs is offered in the List of strong verbs below.

extra
Literature

The traditional division of the strong verbs is along the lines of the original seven Germanic ablaut classes. Since it has become muddled in the later development by all kinds of analogies, this division is given up by modern grammarians. As is also done in the list of strong verbs, they confine themselves to an alphabetical list. Examples are Hoekema (1996:54-69), Tiersma (1999:77-82), Eisma and Popkema (2004) and Popkema (2006:331-342). The traditional division can still be found in Sipma (1913:71-74) and, with direct references to the situation in Old Frisian, in Sipma (1949:31-36). Also with respect to the dialects on the island of Terschelling: Knop (1938). A mixed sytem is offered by Fokkema (1967:70-74). Van Blom (1889:144-154) and Sytstra and Hof (1925:141-147) have extended the division to eleven classes, mainly along the criterion of rhyming.

Kalma (1938:7-8) mentions the strong palatalized form baetsen baked. A recent treatment of the historical development of Frisian strong verbs is offered by Strik (2015).

[+]Inflectional forms of strong verbs

Strong verbs do not deviate from weak verbs in their inflection of the present tense. As for an infinitive ending in -e or -je, they follow the rules as presented in weak verbs for class I and class II weak verbs. Likewise, the forms for the imperative, present participle and infinitive II need no extra attention, as they also follow the rules for weak verbs. Hence, the present tense has the endings -, -st and -t for the three persons singular, and the plural suffix is -e for most strong verbs. For the verb fine to find we then get ik fyn I find, do fynst you find, hy fynt he finds and wy/jimme/hja fine. This is exactly the same pattern that we find with class I verbs. Only a few strong verbs, displaying an infinitive ending in -je, appear to go along with the weak verbs of class II. These are bergje to store, fergje to require, hingje to hang, lykje to appear, terskje to thresh, tingje to haggle and waskje to wash. The present tense of these verbs follows the rules for the weak verbs of class II. To be concrete, for hingje to hang we then get for the present tense: ik hingje I hang, do hingest you hang, hy hinget he hangs and wy/jimme/hja hingje we/you/they hang. Furthermore, the present participle is hingjende and the imperative is hingje.

At first sight, there seems to be a difference for the past tense. Next to the vowel change in the stem, there is the following inflectional pattern:

Table 2
Number Person Suffix
SG -
1 -st
2 -
2 polite -en
3 -
PL
1 -en
2 -en
3 -en
For example, applied to the past tense stem naam (of the strong verb nimme to take), this results in the following forms:

1
a. ik naam
I took
b. do naamst
you took
c. jo namen
you.POL took.PL
you took
d. de man naam
the man took
e. wy namen
we took
f. jimme namen
you.PL took.PL
you took
g. de manlju namen
the men took

The suffixes represent person and number. A comparison reveals, however, that if we abstract away from the past tense markers -e/-de for class I or -e for class II, these endings are similar to what we see with the past tense of weak verbs. Likewise, one may observe the longer suffix -ste for the second person singular, for instance in the form do naamste you took.

It should be noted that the inflectional pattern for the past tense is the same in all cases, irrespective of the infinitive ending in -e or -je. A verb like hingje to hang shows the following preterite forms: ik hong I hung, do hongst you hung, hy hong he hung and wy/jimme/hja hongen we/you/they hung. The past tense suffixes of class II are not added, so we do not have *ik honge or *do hongest or *hy honge. And the past participle is prototypically hongen, and not a *honge or even *hongene.

It is worth mentioning that in the older language (18th/19th century) also a longer past plural ending -ene occurred. Probably, this was the result of a metathesis rule which turned -en to -ne, with the subsequent insertion of a schwa. Relics of this metathesis can still be found in a small set of verbs with a past tense stem ending in a centralizing diphthong, see past tense stems ending in a vowel.

extra
Literature

The existence of the ending -ene was first noticed in Dyk (1996). That such endings are the result of metathesis is argued for Hoekstra (2008).

[+]Irregular weak verbs belonging to class II

Next to the strong verbs, there is a group of verbs displaying different vowels in the past tense and the past participle as well. However, their past participle does not end in -en but rather in a form familar from the weak verbs. In fact, these verbs are weak. They do not belong to the set of traditional strong verbs with ablaut. More specifically, their deviating vowels are the result of historical phonological processes. The two weak classes have their irregular verbs. In class I, shortening has applied. The irregular forms of class II are the result of palatalization of a velar final consonant of the stem.

For the irregular verbs of class II we have to differentiate between stems ending in the voiceless velar stop /k/ and those which have a voiced velar fricative /ɣ/ as final segment. The former, combined with /j/ of the verbal suffix -je, end in a cluster /tsj/ after palatalization. A side effect of this palatalization is that the vowel turns to a diphthong ending in /j/. That change only applies before the suffix -je. This implies that only the infinitive, the first person singular and the three persons plural of the present tense, the present participle and the imperative are affected. For the verb meitsje to make, for example, we thus get the following paradigm for the present and past tense:

Table 3
Tense Number Person Form
present
SG
1 meitsje
2 makkest
2 polite meitsje
3 makket
PL
1 meitsje
2 meitsje
3 meitsje
past
SG
1 makke
2 makkest
2 polite makken
3 makke
PL
1 makken
2 makken
3 makken
In addition, the past participle makke remains unpalatalized. This is not the case for both infinival forms meitsje and meitsjen, the present participle meitsjend(e) and the imperative meitsje!.

The irregular weak verbs with a stem ending in /k/ are the following:

Table 4
Infinitive 3SG.PRET Translation
koaitsje kôke cook; boil
laitsje lake laugh
meitsje makke make
ploaitsje plôke pluck
reitsje rekke;rakke get; hit
smeitsje smakke taste
weitsje wekke watch
The now obsolete verb loaitsje to look could be added to this set. At least for a number of the speakers levelling has occurred to a form with original /k/ in koaitsje, ploaitsje and weitsje. This resulted in the infinitives kôkje (also shortened to kokje), plôkje and its shortened variety plokje, and wekje, respectively.
extra
Dialectical differences

The palatalization of /k/ never took place in the dialect of Hindeloopen. Of the verbs in the table above, Blom (1981) mentions the forms kòkje, lakje, meikje, plòkje, reikje, smeikje and wekje. It may also be the case that this extended to some verbs in other southern dialects. Southwestern forms like kokje, plokje and wekje are possible kandidates. In the Schiermonnikoog dialect, some of the verbs above did not palatalize either, as seen in lakje to laugh and plokje to pluck. In the other verbs, as is generally the case in this dialect, we do not see assibilation. So, verbs like katje to cook or metje to make are without an intermediate /s/.

The irregular verbs of class II with a stem ending in /ɣ/ also underwent palatalization. We may assume that the original sound turned to /j/, which then merged with the initial sound of the suffix -je. Apart from that, the distribution of the palatalized forms is the same as above. One peculiarity, however, is the fact that we see no final schwa in the first person singular of the present tense, in contrast to the normal paradigm of the weak verbs of class II. As an example, here is the paradigm of the verb jeie to hunt.
Table 5
Tense Number Person Form
present
SG
1 jei
2 jagest
2 polite jeie
3 jaget
PL
1 jeie
2 jeie
3 jeie
past
SG
1 jage
2 jagest
2 polite jagen
3 jage
PL
1 jagen
2 jagen
3 jagen
The past participle is jage. We also see palatalized forms in both infinitives (I and II), in the present participle and in the imperative.

The irregular weak verbs with a stem historically ending in /ɣ/ are the following:

Table 6
Infinitive 3SG.PRET Translation
feie fage sweep
jeie jage hunt
kleie klage complain
koaie kôge chew
krije krige get
loeie loege pile
ploeie ploege plough
toaie tôge carry
It should be noted that as a consequence of the latest spelling reform in 1980 it is less conspicuous that these are class II weak verbs; before, the orthography was feije etc.
extra
Other possible members

It remains to be investigated whether there have been more members of this group. A possible candidate is the verb druie to dry, which exists next to droegje. The latter seems to be a direct conversion of the adjective droech dry. The first could be a product of palatalization, although it must be admitted that the verb nowadays belongs to the regular weak class I. A comparable pair is kluie to gnaw next to kloegje, both forms being fairly obsolete. For the tendency of these palatalization verbs to convert to class I, see the section on schwa deletion below.

Another possible candidate is boeie to turn a horse to the right, which might be connected to bûgje to bow, although, according to Dyk (1984-2011) this verb is regularly inflected as a member of the weak class I.

There is quite some levelling within this group. The verbs with a back vowel tend to be restored to a single stem with /ɣ/; we then get the infinitives kôgje to chew, loegje to pile, ploegje to plough and tôgje to carry. The others show levelling in the present tense to a stem only with /j/. For instance, for many speakers the present tense of krije to get reads as follows:
Table 7
Present tense Translation
ik krij I get
do krijst you get
hy krijt he gets
wy/jimme/hja krije we/you/they get
Usually, we do not see levelling of the past tense and the past participle in these verbs: the form with /ɣ/ remains in use. Thus we have ik krige I got and ik haw krige I have got. Further levelling is not excluded, however. For some speakers, verbs like feie to sweep of kleie to complain may have become fully weak. On the other hand, for quite some speakers krije has become strong, with a past tense kriich and a participle krigen. This may possibly have occurred under the influence of the Dutch counterpart, which is also strong.
extra
Schwa deletion

A peculiar feature of this second group of irregular weak verbs of class II is the deletion of the final schwa both in the first person of the present tense and in the imperative. We have ik krij I get and the imperative krij!. However, we do not see *ik krije or *krije!, although these would be the forms expected if the verb follows the paradigm of class II verbs in a regular manner. This could be an indication that the form of the Frisian imperative is rather based on the form of the first person instead of the stem (for this issue, see also categories and also the imperatives of the category of monosyllabic verbs).

A possible explanation might be found in the phonology. In the normal case, class II verbs never occur with a stem ending in a vowel. Instead, we find -je only after a consonant. If, then, the final schwa deletes, this will always result in a final cluster -Cj, which is intolerable in the phonotactics of Frisian. Only here, where /j/ happens to follow a vowel as the accidental effect of palatalization, we happen to meet the suitable context for the deletion of the final schwa.

The circumstance of schwa deletion in the first person of the present tense might have the further effect of levelling in the present tense. The first person form krij might be reinterpreted as a form of the weak class II, which also lacks an ending after the stem. This could be reinforced by the form of the present plural, which is krije. Analyzed as a class II verb, this would be krij-je. However, phonetically there is no objection against the analysis krij-e. The pronunciation is similar, but in the latter case the verb would be a member of class I. If so, it is conceivable that the two remaining forms would also turn to class I. This would result in the second person singular krijst (analyzed as krij-st) and the third person singular krijt (< krij-t), which are the actual forms for many speakers. (For a comparable effect in the realm of conversion verbs, see Hoekstra (1998:152)).

[+]Irregular weak verbs belonging to class I

As stated above, weak verbs of class I may become irregular as a result of a historical shortening in a part of the paradigm. In fact, we see that the non-affected members are subjected to other sound changes, so that the difference within the paradigm even increases. With respect to irregular class I verbs, there are three subtypes. One has a shortening resulting in /ɛ/, another in /a:/ (after a subsequent compensatory lengthening, see Spenter (1968:16)), and finally we have a change resulting in /I/. From a historical perspective the distribution within the paradigm is the opposite of what we saw with the irregular weak verbs of class II. That is, those members of the paradigm that remained stable in class II are typically affected by the sound change in class II. The overall effect is that we find an equal distribution of the allomorphs, but in class I we typically see effects in the second and third person of the present tense, the whole of the past tense and the past participle. Here is a paradigm of the verb bliede to bleed, which belongs to the first subtype:

Table 8
Tense Number Person Form
present
SG
1 blied
2 bletst
3 blet
PL bliede
past
SG
1 blette
2 blettest
3 blette
PL bletten
infinitive I bliede
infinitive II blieden
present participle bliedende
imperative blied
The following verbs belong to this subtype with /ɛ/:
Table 9
Infinitive 3SG.PRET Translation
bliede blette bleed
briede brette roast
fertriette fertrette vex
fiede fette feed
liede lette ring
moete mette meet
rêde rette save
riede rette guess
sliepe slepte sleep
In this group, we find levelling to a high degree, especially in the present tense. The effect is that the short sound /ɛ/ is losing ground in favour of the sound of the infinitive. Levelling can also go in the opposite direction, see infinitives like fertrette to vex and mette to meet in Veen (1984-2011). The forms with /ɛ/ of the verb sliepe are mainly found in the eastern part of the language area; in other parts, breaking may apply, see weak verbs.

For the irregular weak verbs of class I showing long /a:/ as alternating vowel, we take spriede to spread as an example. Here is an abridged paradigm:

Table 10
Tense Number Person Form
present
SG
1 spried
2 spraatst
3 spraat
PL spriede
past
SG
1 sprate
2 spratest
3 sprate
PL