Home ] Up ]

Derivation of nouns


The oldest word form might be the davAr class. They show no sign of syntactical differentiation like the initial-stress-derived hayyal and segholates. That the same roots rarely exist both in davar and in hayyal and segholate forms suggests that most davar nouns succumbed to initial-stress-shift, and either acquired gemination (hayyal) or reduced (segholates).

Because nouns are commonly positioned after verbs, the syntactical accent caused their accent to shift backward,[1] dAvar.

Some initial-stress-derived nouns acquired gemination and formed the hayyal class, davAr - dAvar – dAbbar (post-tonic gemination) – dabbAr (kamatz reduced to patah in closed unaccented syllable). Gemination protected post-tonic vowel against reduction.


Other initial-stress-derived nouns lack gemination. Such nouns became segholate, calAv – cAlav – cal’v.

Vocal schwa lost its sound: calb could be easily pronounced as a single syllable, rather than cal’b. Vocal schwa was lost also in pre-tonic position, cal’bI – calbI.

The Masoretes introduced dagesh to break consonantal cluster, as elsewhere: calv – cal.b; ni-zchar – niz.car.[2]

Epenthetic vocalization came at the expense of the full vowel, calb – celev. Compare cotevt – cotevet. The Masoretes heard epenthesis in chanting.

Segol normally elongates to tzere (talmi’dnu – talmideinu). In the pausal form of segholates, segol elongates to kamatz. That suggests that kamatz was originally there, shortened to segol in epenthesis, calb (kamatz) – celev.

The form c’lavim also suggests the original calav with two kamatz. The second kamatz is preserved in c’lavim. Only kamatz, tzere, and holam reduce to schwa, and the first vowel in c’lavim was plausibly kamatz.

An alternative to gemination is elongation. Both methods protect the post-tonic vowel: either by stop or by softly extinguishing the accented syllable. Ashkenazic accent shift similarly caused elongation, davAr – dAvar – dA:var – dOivor. Elongation does not fully protect post-tonic vowel, and in Ashkenazic second kamatz reduced to short [o] mistaken for holam. Similarly in segholates, cAlav - cA:lav – cElav – cEl’v.

Composite vowels do not survive in heavy syllables, and turn patah when accented, celv – calv. In unaccented syllables, tzere turns short [ae] or hirek. Those segholate words which are generally employed with suffix or constructus, and whose only vowel is thus unaccented, acquired hirek, sefr – sifrI - sifr.

The form kutl similarly appeared as kAtal – kA:tal – kOt’l – kutl.

Plural form retained word-final accent, calav – c’lavIm, and there is no need for dagesh kal.

The constructus plural, c’lavim – c’l’vei (accent shift in constructus) – calvei (patah instead of the more common hirek).


Imperative verbs form nouns with ל, מ, כ, ב prepositions.

Prepositions were originally vocalized, and produce no metathesis and dagesh, c’ch’thov - cich’thov.

Noun-derived infinitives are modeled on the verb form which has unvocalized prefix and produces metathesis, lzacor – liz.cor. The noun form should be l’z’chor – liz’chor. The form l(e)argish also shows that preposition l is vocalized; compare with targish where unvocalized prefix assimilated the word-initial vowel a.


Adding prefix מ to paal created a class of nouns denoting the essence (place or result) of the verb’s action. מcatev has changed into מictav similarly to ncatev – nictav.

Prefix mem could not be straightforwardly applied to other verb stems. All of them produced very similar “nouns of essence”: whether write or dictate, the result is a letter. Mem-nouns from the verb stems other than paal were attached a different meaning: not the action’s result, but its subject or object. מ did not go well with נ, and the participle of nifal lacks מ: nicnas instead of mincanes.

Then, a gap appeared for the paal verbs: they lacked derivative nouns denoting subject of their action (מ-nouns of other verb stems). Such nouns were initial-stress-derived from basic verbs, catEv – cAtev – cA:tev – cOtev.

Subject nouns proved unnecessary, and came to be used as present tense. Syntactical accent of verbs move the accent, cOtev – cotEv. The reduction cotev – cotvim shows that holam is actually short [o], unable to reduce. Pre-tonic vowel is not normally reduced, but the artificial present tense is reduced irregularly.

Inventors of the cotev form also introduced feminine suffix tav instead of the standard, though weak, hey. Tav in cotev was modeled on 2fs past tense tav suffix. Suffix tav could not be pronounced in lamed”hey verbs, and they retained suffix hey. Ambiguous innovation of suffix tav necessitated another deviation, epenthesis cotevt – cotevet instead of *cotvah.


Four-letter nouns with initial א, ת, ה are the future tense verbs.


Suffix –ot refers to uncountable plurality, a multitude forming a new object. Suffix –ut is a reduction of –ot.[3]

Words with suffix –on commonly have dagesh in the last consonant. The only such class is segholate. Likely, -on is the old version of –ot. Suffix -on was employed when segholates were the only class of nouns.

ת commonly replaces ה, and ה replaced נ in the nifal imperative-derived nouns. Suffix –ot is preferable to –on where final nun drops out. Both נ and ת are confirming consonants for the holam suffix.

Similarly to suffix –ot, -on denotes abstract plurality, Helbon – place of milk. –un suffix is a reduction of –on.

The noun classes should not be distinguished based on their vowels. Rather, two classes should only be regarded as different if some of the same roots are encountered in both. By this definition, -on, -un, -ot, and –ut suffixes form a single class.


Nouns of shomer and shamea classes are semantically evolved participles.

Nouns of d’vash class are davar nouns generally employed in constructus which reduced the first vowel.

Nouns typically employed in constructus are reduced irregularly: karev – kar’vA – k’r’vA (constructus) – kirva.

Some verbs of catev class, employed as nouns, lost syntactical stress, barEh – barIh.


Adjectives derive from nouns, adam – adom. Usage of nouns as adjectives is modeled on constructus, where the second word is semantically similar to adjective. Second words of constructus have strong syntactical accent, Adam – A:d’m – A:dom with short [o].

Other adjectives evolved from imperative or future tense, gadOl.

Adjectives were standardized around the gadOl form, and [o] in adOm became long.

Derivative nouns describe the result (“destination”, thus ה) of the quality: gadOl – gadOlla (penultimate accent of nouns) – g’dullA (unaccented holam turns shuruk).

Feminine gdolA (instead of gdulA) was the only choice to distinguish from passive participle ctuva, noun gdulla, and 3fs past tense catva.

[1] The stress shift is evident in Germanic. In English, both verbs and nouns borrowed from Germanic were pronounced with the same accent. Eventually, the phonetic tendency of differentiating stress in verbs and nouns based on intonation shifted the accent in English, too, recOrd - rEcord.

[2] Segholates might appear because of the suffix shuruk, calAbu – calAbbu (post-tonic gemination) – calAbb (unaccented final vowel dropped) – cAlabb (initial stress shift in nouns) – cAl’bb (two consequential geminations would be awkward) – calb (loss of the schwa vocalization - dagesh kal breaks cluster). Plurals lacked suffix shuruk, and produced no gemination (dagesh).

[3] Suffix –ut could appear by mistake. Conjugated imperative of paal ל"ה, zanoה – zanot –znut. The imperative-derived nouns from ל"ה root often refer to abstract notion, possibly related to the semantics of ה. The final –ut was taken for suffix and applied to create other nouns.