The Endless Story of “Compound Verbs”: An Introduction to Conceptology and Terminology of Complex Verbs in Persian

Much of the literature about Persian ‘compound verbs’ has been concerned with their ‘true’ instances. About four centuries after identification of this construction, however, there is yet little agreement on any single definition or conception of it among scholars. The present article aimed to show that this controversy was primarily a matter of terminology rather than conceptology. To this end, after reviewing major authors’ views on so-called "compound verbs", we examined various kinds of constructions involved according to linguists’ received definitions of the two concepts of ‘verb’ and ‘compound’. It turned out that i) there was no good reason for regarding ‘compound verbs’ as verbs and ii) being separable strings, they could hardly be considered as proper morphological “compounds”. We then suggested that all these kinds of constructions could well be subsumed under the rubric ‘complex predicates’ (whether semantic or grammatical), thus making it more likely to reach a consensus on the issue. 

The commonalities of meta-analytical literature on “compound verbs” in Persian can be summarized as follows:Taking for granted the existence of ‘compound verbs’
Examining and evaluating the current definitions of ‘compound verbs’ according to their own conceptions
Endorsing one or more previous opinions (with some minor modifications) without dealing with the nature of disagreement among scholars in recognizing the instances of ‘compound verbs’.

The variety of strings referred to as “compound verbs” covers a vast range. Having reviewed and compared representative opinions of authors and their definitions of these strings, we can sum up their viewpoints as follows:Incorporational verbs (like γazāxordan ‘to eat food’, lit. ‘food eat’) and verbal phrases (like azdastdādan ‘to lose’, lit. ‘from hand give’) were regarded as ‘compound verbs’ only by Dabir-Moghaddam, and experiencer-subject LVCs[1] (like dard-am gereft ‘it pained me’, lit. ‘to me pain took’) only by Khanlari.
Dabir-Moghaddam and Abolghassemi considered (some of) the auxiliary verb constructions (present perfect, past perfect, and passive) to be ‘compound verbs’.
Khayyampour and Dabir-Moghaddam regarded prefixed verbs (like bar-dāštan ‘pick up’) as ‘compound verbs’.
Vahidian did not believe in ‘compound verbs’ to exist in Persian at all.

Looking at the above-mentioned definitions by the authors, it was noted that their conceptions of ‘compound’ simply meant an expression that was not simple, i.e., not having more than one element, which is usually called ‘complex’ (as opposed to ‘simplex’) in English. A ‘complex verb’ is a verb with more than one element. The additional elements could be the result of various kinds of affixation (including personal endings or other inflectional or derivational affixes) or compounding. A proper ‘compound verb’, however, is a verb that is formed specifically through the morphological process of compounding, in which two lexical items (typically verbal roots) join together to form a lexeme.
Therefore, if by ‘compound’ we mean ‘complex’, i.e. by ‘compound’ consisting of two or more elements, then, a ‘compound verb’ would be certainly complex. Terminologically speaking, however, morphological criteria like inextensibility could not be excluded from compounds, including alleged ‘compound verbs’, which do not meet such criteria. For example, although kār(hā) kard ‘(s)he did thing(s)’, lit. ‘work(s) did’ despite its extensibility, is undoubtedly complex, to regard it as a morphological compound requires us to revise the concept of ‘compound’.
As for ‘verbness’ of a ‘compound verb’, given the linguists’ received definition of a verb, according to which ‘verb’ is the locus of indicating tense and often aspect, mood, and agreement, ‘verbness’ of so-called ‘compound verb’ was highly arguable because most of them were split by inflectional affixes intervening their two components. Thus, before referring to any of these strings as a ‘verb’, we needed to make clear in which sense their preverbal elements were parts of the verb.
It seemed that what the authors, such as Khanlari and Dabir-Moghaddam, meant by ‘verb’ in a ‘compound verb’ was a ‘predicate’ rather than a verb as such. This terminological confusion was rooted in the term ‘predicate’ having at least the following two senses:In semantics and formal logic, it is an element of a proposition, which express some property of, or relation between, argument(s). This kind of predicate can be called ‘semantic predicate’.
In traditional grammar, it is a part of a proposition, which does not include the ‘subject’. This type of predicate we can refer to as ‘grammatical predicate’.
Based on what we said about the terms ‘verb’ and ‘predicate’, it should be clear that what was referred to as ‘compound verbs’ in the literature was in fact a (semantic) complex predicates of the clauses involved. Thus, regarding them as ‘verbs’ as mentioned was a result of i) a confusion between the terms ‘verb’ and ‘predicate’ and ii) the fact that (semantic) predicates are typically verbs, while the verbs often consist of predicates of the sentences, in which they occurr.
To conclude, different kinds of strings, only a few of which were considered ‘compound verbs’ by some authors (and for which we questioned this designation), could well be included under rubric ‘complex predicates’ in spite of their structural variety.