The paradigm of Greek verbs is divided into perfective and imperfective forms (or, in traditional terminology, aorist stem and present stem), an opposition both of aspectual semantics and morphological stem. Verbs which inflect for only one aspect are traditionally termed defective. Most of these are restricted to the imperfective. The list given by Holton, Mackridge and Philippaki-Warburton (1997: 119) gives the impression that this restriction is at least in part due to the verbs' semantics, since most of them have a stative or durative meaning, e.g. aníko 'belong', apekhthánome 'loathe', asthméno 'pant, be breathless', vrítho 'be full (of)', jitonévo 'be in the neighborhood of', diejíro 'excite', diépo 'govern', íme 'be', érpo 'crawl', evthínome 'be responsible', ékho 'have', lámno 'row', mákhome 'fight', méllo 'be about to/destined to', ksero 'know', odírome '(be)wail', ofílo 'owe, be obliged', perimméno 'wait (for)', repo 'incline', tremo 'tremble', khásko 'gape' and khrostó 'owe'.
A much smaller group of verbs have only perfective forms, e.g. koréstika 'I was stuffed, satiated', khrimástisa 'I served [as mayor etc.]' (Adams 1987: 76).
The active and passive past participles are learned, literary form, of limited both in their use and in the range of verbs that form them. What is of interest here is not the restrictions on the participles as such, but the restrictions on the paradigms of those participles that are in use. Their archaizing inflection is aberrant. In the case of the past active participles, there is at least one type where only a few forms out of the total conceivable paradigm are ever used (Holton et al. 1997: 166):
| 'writer' (lit. 'the having written one') | |||||
| M SG | F SG | M PL | F PL | ||
| NOM | grápsas | grápsasa | grápsantes | ------ | |
| ACC | grápsanta | ------ | ------ | ------ | |
| GEN | grápsantos | ------ |
grapsánton
|
||
The verbs theto 'place' and -isto (a bound stem), which host a large number of prefixes, have defective or problematic passive forms. While the active forms are inflected normally, the passive has, in part, archaic endings. Compare the following forms of the irregular theto and -isto (here kath-isto 'render') with those of a regular verb:1
IPFV PRS PASS PFV PST PASS IPFV PRS PASS PFV PST PASS IPFV PRS PASS PFV PST PASS 1SG tith-eme tethik-a kathist-ame katest-in graf-ome graftik-a 2SG tith-ese tethik-es kathist-ase katest-is graf-ese graftik-es 3SG tith-ete tethik-e kathist-ate katest-i graf-ete graftik-e 1PL (tith-emetha) tethik-ame ------------ ------------ graf-omaste graftik-ame 2PL (tith-esthe) tethik-ate ------------ ------------ graf-este graftik-ate 3PL tith-ente tethik-an kathist-ante katest-isan graf-onte graftik-an
The singular and 3rd plural forms of theto and kathisto are close or identical to those of a regular verb. The 1st and 2nd plural of theto are highly irregular, and correspond to the Ancient Greek endings; Holton et al. (1997) note that they are little used, while kathisto lacks these forms altogether. In the perfective past, theto has regular forms while kathisto is wholly irregular, the forms once again being those of Ancient Greek.
While the problems with the 1st and 2nd plural in the imperfective present passive can be attributed to the irregularity of the (presumably) expected forms, the reason for the defectiveness of the perfective past passive of kathisto is less clear, since the whole paradigm is irregular anyhow. Perhaps it can be seen as an echo of the gap in the imperfective present passive.
Sims (2006) describes how the genitive plural of many nouns is avoided (the number is quite large, well over 1,000 according to dictionaries). Note that in most contexts, there is variation between the genitive and prepositional phrases with the accusative which are functionally equivalent, even for nouns whose genitive plurals are not avoided.
These gaps are clearly related to the exceptional status of the genitive plural in the accentual system. Greek nouns fall into various declension classes, and stress placement is, to a certain degree, predictable from the declension class. The declension classes fall into three types:
Type 1: Stress is fixed on the same part of the word for the entire paradigm.
Type 2: Stress shifts in the course of inflection. All the stress shifts involve the genitive plural:
|
b. shift in GEN
'tourist' |
d. shift in PL
'power' |
|||||||
| singular | plural | singular | plural | singular | plural | singular | plural | |
| NOM | níki | níkes | aγóri | aγórja | ípiros | ípiri | ðínami | ðinámis |
| ACC | níki | níkes | aγóri | aγórja | ípiro | ipírus | ðínami | ðinámis |
| GEN | níkis | nikón | aγorú | aγorjón | ipíru | ipíron | ðínamis | ðinámeon |
Type 3: For these declension classes whether or not they undergo a stress shift must be lexically specified. Nouns belonging to these classes behave either like Type 1, or like Type 2 (a, b or c).
The bulk of nouns for which the genitive plural is avoided belong to Type 3, and are relatively infrequent, suggesting that uncertainty about the stress placement in the genitive plural plays a role. But, Sims argues, there is no evidence that uncertainty is the direct synchronic motivation. In many cases the gaps appear simply to be lexically specified.
1 The forms of rest of the paradigm of theto and -isto are essentially regular, but note that in the imperfective past passive both verb stems are said to be restricted to the 3rd person.
Adams, Douglas Q. 1987. Essential Modern Greek Grammar. New York: Dover.
Holton, David, Peter Mackridge, and Irene Philippaki-Warburton. 1997. Greek: a comprehensive grammar of the modern language. London: Routledge.
Sims, Andrea. 2006. Minding the gaps: inflectional defectiveness in a paradigmatic theory. PhD thesis, The Ohio State University.