The (Lack of) Evidence for the Qal Passive in Biblical Hebrew

There is widespread agreement among scholars that Biblical Hebrew (BH) has a Qal passive which was subsequently obscured by the Massoretic vocalisation. Waltke & O'Connor¹ suggest there are some 160 verb forms which are probably Qal passive, rising to nearly 200 if participles are included. The version of the parsed Hebrew Old Testament that I have lists a more conservative 81 occurrences.

The textual argument has a number of components. Firstly, it is argued that some suffix conjugation (SC) forms vocalised as Pual are actually Qal passive, as the Piel/Hithpael is unattested, has a different sense, or the Piel occurs less frequently than the Pual. Likewise, some prefix conjugation (PC) forms are vocalised as Hophal, but Hiphil forms are unattested or have a different sense. This leads to some verbs having SC Pual complementing PC Hophal. Further examples of Qal passive SC forms can be found improperly vocalised as Niphal PC forms. Finally, some participles have forms that are improper for Niphal, Pual or Hophal.

It is also noted that cognate languages such as Ugaritic and Arabic have internal passives, which are formed by vowel changes to the consonantally distinct verb stems.

Overwhelming as the evidence might seem at first glance, I suggest that there is actually little if any real evidence for the existence of a Qal passive in Biblical Hebrew.

The evidence for Qal passive pointed as Suffix Conjugation Pual

At first sight the relative superabundance of Pual SC may appear to lend support for the existence of a Qal passive. However, the semantics of Pual are critical here. Waltke & O'Connor develop an idea first proposed by Jenni², that the Pual (and Piel) emphasise the result of an action rather than the action itself, i.e. loosely speaking, that they are resultatives. With this understanding of the Pual, its distribution is unremarkable. One of the verbs occurring most frequently in Pual is yld 'to give birth'. With this verb, Niphal is appropriate when indicating the birth itself (usually during reference to its temporal or spatial location), whereas Pual focusses on the living being resulting from the birth. Pual is very frequent in genealogies where the sequence of individuals is of prime interest, not the actual births themselves. When we talk about a present state (being alive), the birth that led to it lies in the past, hence the exclusive use of SC Pual for this particular verb³. There are likewise numerous Pual SC forms of šdd 'to devastate', but once again, in each case, this can be understood as emphasising the resultant desolate state of the land. Similar arguments may be made for all the other "suspect" Pual SC forms cited by Waltke & O'Connor.

The evidence for Qal passive pointed as Prefix Conjugation Hophal

Waltke & O'Connor list 28 verbs pointed Hophal which, primarily on semantic grounds, might be considered to be Qal passives. All these verbs have roots which are either geminate, nun initial (including the anomalous lqx), or yodh or waw medial. However, there is no reason why the use of the Qal passive should have any connection with the morphology of the verb root! Verbs in the above categories occur in total a little under 18,000 times in the OT corpus, whereas regular 'strong' verbs occur about 22,000 times. Treating these as approximately equal, simple binomial theory suggests that the probability of all the putative Qal passives occurring in one group, and none in the other, is 1 in 2²⁷, or less than 1 in 130 million. Even if half the proffered examples are actually Hophal, the probability of the observed partitioning of the remainder is still 1 in 2¹³, or less than 1 in 8,000. In other words, whatever the reason is for the semantically anomalous Hophal forms, they cannot be Qal passives, as indicated by the above statistics. What they actually do represent is rather beyond the scope of this post. However, there are apparently semantically identical verbs that occur in paired AwB and yAB forms for at least ṭwb/yṭb and gwr/ygr, and according to my parsed OT, also for mwr/ymr, ʿwp/yʿp, ʿwc/yʿc, cwq/ycq and cwr/ycr. Many of the suspect Hophal for waw/yodh medial verbs could be Pual if a yAB variant exists, and similarly for the geminates if ABB forms have yAB variants. The bulk of the suspect Hophal for nun initial verbs are for ntn and lqx (which is irregular, but behaves like a nun initial verb). For neither verb do we have Piel forms for comparison, but it is possible that the initial nun or lamedh is dropped for these particular verbs to give Pual forms identical to Hophal, i.e. yənuttan → yuttan and yəluqqaḥ → yuqqaḥ.

The evidence from cognate languages

In both Arabic and (apparently) Ugaritic, any verb stem can potentially have internal passive forms, restricted only by the semantics of the active form yielding a meaningful passive. In the BH binyanim system, the Piel, Hiphil and Hithpael stems already have passives in the Pual, Hophal and (rather marginal) Hothpael, which only differ in the unpointed text from the corresponding active forms with respect to long vowels marked by yodh and waw, and hence are essentially internal passives. Whilst Niphal is a full stem of a different form to Qal, and has wider uses than just as the passive of Qal, it nonetheless fits into this pattern of paired active/passive forms. A Qal passive thus seems redundant, and judging by the distribution of Niphal and putative Qal passive forms throughout the OT corpus, one is not obviously in the process of supplanting the other.

Other anomalous forms

Besides the two verb conjugations above, a number of anomalous participles are found which could be considered as semantically Qal passive, and which have, or could have, vocalisation appropriate for the participle of a Qal passive form. The primary offenders are yld with yillōd occurring five times and yûllād twice, and škr 'to be drunk' with šikkōr appearing thirteen times. However given the weakness of the evidence for finite Qal passives, and the wide variety of anomalous adjectival forms found in the OT, it is hard to draw conclusions from the existence of such apparent Qal passive participle forms.

I have not dealt with every shred of evidence for the Qal passive here, but the prevalence of Pual SC forms has been shown to relate to the semantics of the Pual form, and the anomalous Hophal forms have been ruled out statistically, as they are associated with particular verb morphology, not semantics. The remaining peculiarities that might suggest the presence of the Qal passive are sufficiently few and far between to be indistinguishable from the background level of noise from errors in the unpointed OT text and inconsistencies in its subsequent Massoretic pointing. Of course I cannot be certain that the Qal passive does not exist, only that the evidence usually adduced to support it does not do so to any significant degree...

---

¹B. Waltke & M. O’Connor, An Introduction to Biblical Hebrew Syntax (Winona Lake: Eisenbrauns, 1990)
²E. Jenni, Das Hebräische Piʿel: Syntaktisch-semasiologische Untersuchung einer Verbalform im Alten Testament (Zurich: EVZ, 1968)
³Distinguishing PC Niphal and Pual forms of yld is potentially problematic as the underlying verb root wld might lead to yəwullad → yûlad with the lamedh losing its dagesh as the preceding syllable is now open with a long vowel. Such a form when unpointed would be identical to the Niphal, and hence some of the many apparent PC Niphal forms for such waw initial verbs in the OT may have been pointed incorrectly by the Massoretes.