Middle Voice, Deponency & Language Typology

One of the questions put forward to the panel in the discussion was on the lack of comment from any of the panelists about Suzanne Kemmer’s book The Middle Voice (Typological Studies in Language 23). Stanley Porter responded to the question. He basically dismissed the use of typological study as being unhelpful and potentially dangerous for studying Ancient Greek grammar.

My jaw dropped.

Now, let me up front say that while Porter and I have our differences on a couple issues (e.g. tense vs. remoteness), I hold much of his work (especially his recent “introductory” grammar) in high esteem. Take the section on pronouns for example: he has provided students and scholars with the best description of accented pronouns (non-clitic) in existence. That’s saying something because there are plenty of grammars out there. I ask my reader to keep that in mind while reading my following statements.

To speak frankly, Porter is dead wrong on typology.

Typology is risky and dangerous for studying Greek grammar. Why? Porter basically stated that using typological studies is dangerous because they have the potential to mislead the scholar to draw conclusions about Greek grammar not from the internal structure of the language, but instead the structure of other languages.

There is always a danger of the scholar (generally unintentionally) making the language they’re studying look more like or less like other languages. But this is a danger that is completely separate from typological research. In fact, I would suggest that the danger being realized is more likely when someone is exposed to one or two languages rather than when one is exposed to many languages. For the English speaker who is studying Ancient Greek grammar, that person is going to either be inclined to make Greek more like English or less like English either consciously or unconsciously. Its going to happen. If anything, being familiar with what a broader set of  languages do is only going to function as a safe guard against these two inclinations rather than making it more likely.

It wasn’t typological study that led A. T. Robertson in Grammar of the Greek New Testament in Light of the Historical Research to put forward an eight case system for Koine Greek.


It was the study of one language: Sanskrit.

It wasn’t typological study of language that led the grammarians of the centuries before Moulton, Blass, & Robertson to provide a primarily tense-based description of the Greek verbal system.


It was those grammarian’s knowledge of their native tongues and the language of the academy that caused it: Latin, English, German.

If anything, typological/cross-linguistic research saved Greek grammar from a primarily tense based system. It was the comparative grammar on aspect in the 19th and 20th centuries that led to the aspect-prominence descriptions of Greek we have today. In fact, Porter, Fanning, and Decker all thoroughly rely on the typological literature for their descriptions of aspect in their respective dissertations. And I’m not sure whether any of their descriptions would have been possible at all if it hadn’t been for typological research. The irony there is that Decker was nodding his head in agreement with Porter as he made these statements.

Part of me wonders if Porter’s rejection of typological research arises from his own methodological foundation: Systemic Functional Grammar. This framework in which he grounds all of his research does virtually nothing with language typology. Christopher S. Bulter, a former SFG practitioner who abandoned the framework because of its failure to respond productively to criticisms of its flaws writes in his book Structure and Function: A Guide to Three Major Structural-Functional Theories Part I: Approaches to the simplex clause (Studies in Language Companion Series, 63),

Typological concerns have not played a major part in the development of SFG.  For many years, almost all of the central figures in SFG worked on English, though it should be remembered that Halliday’s initial work (1956, 1959) was on Chinese. Three of the most authoritative accounts of SFG (Halliday 1994b; Matthiessen 1995; Martin 1992a) are based very largely on English, though occasional comments on other languages. As we shall see at various points in the present book, these facts have had important effects on the form of ‘mainstream’ systemic linguistics, despite the fact that Halliday himself has warned of the need for care to be taken in this area:

Modern linguistics, with its universalist ideology, has been distressingly ethnocentric, making all other languages look like imperfect copies of English.
(Halliday 1994b: xxxi)

There has been more work on languages other than English in recent years, though we shall see that the stance taken on typological matters is very different from that of FG [Functional Grammar] or RRG [Role & Reference Grammar] (47-48).

At its most basic level, SFG is not a framework that does typological work and this reality is reflected in Porter for better or worse–and I would suggest that it is for worse.

And typological research goes in both directions, too. Typologists beg for high quality descriptive grammars. Thus, Kemmer writes in her preface,

“[L]anguage particular studies are an indispensible supplement to the kind of broad cross-linguistic investigation undertaken here. First, they provide an empirical basis for testing some of the cross-linguistic generalizations proposed; but more importantly, they are essential for arriving at an understanding of the ways that particular middle voice systems interact with other semantic/grammatical systems such as aspect” (x).

If the typological literature is unreliable, it is because we have failed to provide quality descriptions for its use.

And when we look at Kemmer’s work, I am even more at a loss as to why we should ignore it. Ancient Greek is one of the languages that plays a central role in her presentation based on the grammar of Smyth. Porter in a mere handful of sentences dismissed a piece of secondary literature that deals specifically with Greek???

So what is the point of typological study? What is its strength; its benefit?

D. N. S. Bhat in The Prominence of Tense, Aspect and Mood (Studies in Language Companion Series 49)alt gives a helpful discussion here.

We may compare the distinction between typological studies of the above type [large scale, broad coverage studies] and in-depth studies of individual languages with the distinction between areal or satellite pictures of a countryside and an architect’s drawing of a town or a dam. A satellite picture would only show patches of colour and vague lines and curves that an expert can interpret as indicating the location of a possible earthquake or deposits of mineral wealth, whereas an architect’s drawings would show locations of various buildings, parks, canals, etc. in very precise terms.

It would be a mistake, however, to discard the former merely on the basis of the fact that they are not as precise and specific as the latter. The two complement one another, with the satellite pictures giving a warning to the builder of dams so that they can avoid certain cites [sic] as possible disaster areas.  Typological studies and in-depth studies of language can also complement one another in a similar fashion.  The general tendencies that one can perceive through a comparison of hundreds of different langauges can be helpful in avoiding certain conclusions and in raising certain questions that might not have been raised otherwise while carrying out in-depth studies of individual languages.

The relevance of this typological classification for our understanding of the nature of different languages can be perceived even in the very process of carrying out this study. Notice that much of the grammatical information that we will be using as the basis of this study has been elicited and described with the help of one single language, namely English. Since this language is a tense-prominent language, and since, as I have suggested earlier, tense-prominent languages view concepts belonging to other verbal categories in terms of the category of tense, much of our data on aspect and mood would be biased by a temporal point of view. It is something like trying to understand the colour of various objects around us while looking at them through a red-coloured glass.

There are several instances in which a language that was described by a series of grammarians as showing a primary tense distinction, like past, present, and future, has been shown to be actually making a primary aspect or mood distinction. … What is interesting to note, in these, cases, is that the majority of such reinterpretations, as earlier tense-based description had to be rewritten as aspect-based or mood-based description. That is, the need to change an earlier description has occurred in almost all cases as a need to remove the bias that has resulted from our use of a tense-prominent language as the language of elicitation and description (98-100).

Nobody is claiming that typological study is perfect. In fact, I would go as far to say that there is a need for training students how to properly use and evaluate typological & cross-linguistic research. But it is valuable and important for both the avoidance of native language bias and for giving an accurate picture of what is possible in human language. It cannot be ignored or dismissed as easily as Dr. Porter’s comments in the deponency panel discussion might have suggested.