https://d2ng68cf36kxsk.cloudfront.net/store/public/post/686/post_images/Fig.%202.1.2a.jpg">

Fig. 2.1.2a: Transactivity

All three fragments of speech in this example – “Gretl opened the door,” “the door opened,” and “the door was opened” – are “indicative” in their mood. In Wittgenstein’s rather unsatisfactory terms, these are the kinds of propositions we ideally use to picture reality. In the Tractatus, Wittgenstein does not get much beyond declarative propositions of fact, and even more narrowly, he prefers for his facticity active and transitive versions of indicative propositions. He does not seem interested to distinguish these kinds of indicative proposition,^§MS2.1.1d so neglecting large stretches of the universe of agency that we want to capture under the concept of transactivity.

The first of the three fragments is a proposition that Wittgenstein would readily be able to picture. “Gretl,” the subject/theme, “opened the door,” something we have already called predication.^§MS2.1.1 The subject of the statement (a noun in the nominative case, the causer of the action and its subsequent impact on another entity) is marked by tonic emphasis – in this transcription, the bolded words. The predicate includes both a verb referencing action and a noun referencing what has been acted upon (the object, with the noun in the accusative case). For its having a “direct object,” an explicit accusative, the sentence is taken to be in the active voice.

We are not going to look at the second speech fragment quite yet because it is the most problematic. For the moment, we are going to move to the third of the fragments, the passive voice. We have changed the subject, from Gretl to the door. The door was opened by someone, it seems – Gretl perhaps, and we might be able to assume this from the context of the speaking or the situation. In this case, we could have said, “the door was opened by Gretl,” but as it happened, we didn’t. There is no direct object, no accusative, and the door mysteriously becomes the agent. Or perhaps there is a mysterious agent behind the agent.

Speech has allowed us to make a statement where we have left something out, meaningfully for the sake of economy where context clearly speaks, or just as meaningfully because we would rather not say. Speech allows us to commit omission, for one reason or another, and the reason is never immediately clear. The omission may be a matter of redundancy within speech (it’s a waste of time to say it again because it’s already been said, and it may be inelegant to repeat the obvious), or pragmatics (we share the same context, so it should be obvious from that, even if it is not obvious from the speech), or ignorance, or malice. These are all forms of (non-)meaning-by-ellipsis, or omission.

Now we come to the second and most vexing of the three fragments of speech, “the door opened.” Here, “open” has become an intransitive verb, or so it seems, in defiance of the seemingly neat grammatical distinction between one kind of verb and another. In the case of transitive verbs, action involving one entity extends to another (Gretl and the door). In the case of intransitive verbs, the action does not extend to another entity. Something just happens. This is what appears to be the situation when “the door opened.”

The suggestion is that there is no agent. However, here arises another type of agency, called the ergative,^§MS2.1.2d mostly found in speech when a verb can either be transitive or intransitive and the nouns in subject and object share the same case. In fact, this presents a third option. (Halliday calls it “middle,” between transitive and intransitive.) Instead of a straightforward actor, action, and goal where the action is represented in a transitive verb, we have a different kind of agency, that of cause and effect, though the causer is omitted by a process of ellipsis in this fragment. However, there must have been a causer. If not Gretl, then perhaps a gust of wind blows the door open, or the way the door was hung on its hinges means that it seems to open itself. This kind of ellipsis is different from the ellipsis that occurs in the passive voice.

So, even these simplest of actions give us trouble in our account of agency. Fillmore focuses his analysis on the grammatical case of the entities: the “ergative” case is neither nominative nor accusative. It is marked overtly in some languages, but not English. Equally troublesome in terms of slippages between cases, Fillmore points out that the ergative case is frequently connected in English to what would at first glance be considered the simplest of verbs, including “break,” “connect,” “continue,” “develop,” “increase,” and many others.⁸⁶

Halliday, in his analysis, focuses more on the transitivity of verbs than the case of the nouns to which these verbs are connected, where transitive verbs link action and goal, but the ergative links cause to effect.

The operative/receptive forms then contain a second, optional participant whose function might be labelled, for purposes of the discussion, the “causer”; the causer is the subject in the operative and the agent in the receptive, John in both John opened the door and the door was opened by John. The causer is present in the operative/receptive but not in the middle form; in other words, there is an opposition of features middle/non-middle such that the non-middle is interpretable as embodying external causation, the existence of a causer that is not identical with, or at least is treated as discrete from, the affected.⁸⁷

We use the ergative in speech because the cause is unknown, or not wanting to be disclosed, or because the causer is too complex a mix of natural and human causes to bear the immediate trouble of representation. Behind the wind is a window left open by someone, Gretl perhaps? Behind the door that swings open by itself when unlatched is a tradesperson who made it that way, perhaps, or someone who had been working with Ludwig who designed it that way, perhaps? The “perhaps” is the ellipsis.

The ergative refers to cause but buries it, so avoiding having to account for it. But cause nevertheless we know there must be, natural and/or human. Behind every entity is a lineage of action, Alfred North Whitehead would remind us.^§MS1.2.2c To this, we would now add that there must be agency, natural and/or human, whether or not this is disclosed in particular frames of partiality by one pattern of transactivity or another.

Fillmore’s solution to the lack of conceptual clarity in conventional grammars of speech is to suggest a radically different configuration of cases. He suggests, in the interim at least, agentive (animate instigator), instrumental (inanimate force), dative (animate being affected), factitive (object affected), locative (place where affected), objective (things affected by an action). And, he says, “additional cases will surely be needed.”⁸⁸ His conclusion: theories of grammatical case capture the relations of nouns to verbs poorly.⁸⁹

Halliday, for his part, proposes a different system of transitivity where agency is expressed in six clause types: “extensive clause with ‘action’ process-type; effective clause with ‘directed action’ process-type; operative clause with ‘directed action,’ subject as actor; receptive clause with ‘directed action,’ subject as goal; descriptive clause with ‘non-directed action’ process-type; intensive clause with ‘ascription’ process-type.”

“Transitivity,” he says, “is thus being defined in terms of paradigmatic and syntagmatic relations in the clause, not by the classification of verbs as ‘transitive’ or ‘intransitive.’”⁹⁰ Nor, we should add, by the grammatical case of nouns – though of course verbs and nouns in speech work together to create these larger, intrinsically relational units of meaning.

Something happened with that door, if we care to trust any of the three fragments we presented in the figure at the beginning of this section. Let’s assume the same matter of happening to which each fragment refers. However, each configures agency in its own way. Each has its own mode of revelation and occlusion. Such meaning-selectivity Halliday calls “construal.” Each, by transposition, could easily slip into the other and still be true to door-ness and opening, meaning the same thing but slightly differently.

Or, to use the language we have been developing for our grammar, between the door opening (itself a design of meaning) and the fragments of speech that are its transposition, there has been a redesign according to the affordances of speech, necessarily limiting meaning. This involves acts of omission of one kind or another, explicable according to the pragmatics of representational economy, or ignorance, or deception. Without multimodal parsing, it is hard to know the reasons for these ellipses, as well as the dynamic possibilities for meaning that have been elided. The meaning is in the immanent elisions, and their imminent revisability.

So we must say in reply to Wittgenstein^§MS2.1.1a that there is much more to my meanings than my language, and the limits of the affordances of my language are consciously or unconsciously leveraged for reasons that language itself often does not disclose.

To get beyond these limits of language, Fillmore proposed a “frame semantics.”⁹¹ Late in his career, he and several colleagues started to turn their theory of frames into a semantic markup language,⁹² expressed in Standard Generalized Markup Language (SGML),^§AS1.4.7d but the project was left incomplete with the rise of primarily statistical approaches to mechanical language processing.^§AS2.2.1a Frame semantics was intended to be a program by means of which the limitations of language could be supplemented. Such a project still waits to be addressed in the era of computer-mediated meaning.

Language does not readily confess its unremarked features, because ellipsis is one of its key strategies, pragmatically to avoid ponderous over-explicitness, or as a cover for referential laziness, or as a sign of ignorance, or as a matter of deceit. We need a wider examination of frames of meaning to unveil these (non-)meanings.

A multimodal grammar always needs to move beyond speech to give a fuller account of meaning. Multimodal transposition can fill in the dots of the ellipses that are an intrinsic affordance of speech.

---

⁸⁶ Charles J. Fillmore, 1966, “Toward a Modern Theory of Case.” Columbus, OH: Ohio State University, pp. 5–7.

⁸⁷ M.A.K. Halliday, 1968, “Notes on Transitivity and Theme in English: Part 3,” Journal of Linguistics 4(2):179–215, pp. 182, 185.

⁸⁸ Charles J. Fillmore, 1968, “The Case for Case,” pp. 1–88 in Universals in Linguistic Theory, edited by E. Bach and R.T. Harms, New York, NY: Holt, Rinehart and Winston, pp. 24–25.

⁸⁹ Charles J. Fillmore, 1977, “The Case for Case Reopened,” Syntax and Semantics 8:59–82, pp. 70–71.

⁹⁰ M.A.K. Halliday, 1967b, “Notes on Transitivity and Theme in English: Part 1,” Journal of Linguistics 3(1):37–81, pp. 40, 52.

⁹¹ Charles J. Fillmore, 1976, “Frame Semantics and the Nature of Language,” Annals of the New York Academy of Sciences 280(1):20–32.

⁹² Collin F. Baker, Charles J. Fillmore, and John B. Lowe, 1998, “The Berkeley Framenet Project,” paper presented at the COLING ’98 Proceedings of the 17th International Conference on Computational Linguistics, August 10–14, Montreal.

---

Cope, Bill, and Mary Kalantzis. 2020. Making Sense: Reference, Agency and Structure in a Grammar of Multimodal Meaning. Cambridge UK: Cambridge University Press, pp. 198-202. https://doi.org/10.1017/9781316459645

Charles Fillmore on Case §MS2.1.2a