LINGUIST List 32.2374

Wed Jul 14 2021

Review: Cognitive Science; Ling & Literature; Philosophy of Language; Psycholinguistics: Zacks (2020)

Editor for this issue: Jeremy Coburn <>

Date: 08-Jan-2021
From: Marta Donazzan <>
Subject: Ten Lectures on the Representation of Events in Language, Perception, Memory, and Action Control
E-mail this message to a friend

Discuss this message

Book announced at

AUTHOR: Jeffrey M. Zacks
TITLE: Ten Lectures on the Representation of Events in Language, Perception, Memory, and Action Control
SERIES TITLE: Distinguished Lectures in Cognitive Linguistics
YEAR: 2020

REVIEWER: Marta Donazzan, Laboratoire de Linguistique de Nantes (UMR 6310)


This book is the collection of the ten lectures that Jeffery Zacks gave in December 2017 as the forum speaker at the “17th China International Forum on Cognitive Linguistics”in Beijing. It consists of a transcript of the lectures, with no audio-visual material included. Original audio-recordings and supplementary material have been made available online and may be accessed on the website via a QR code. The book, and its associated online material, are a good introduction for all students and scholars specialising in linguistics and interested in understanding the interface between linguistic and cognitive representations. Although the format of the book – a collection of lectures – may sometimes give rise to a certain incongruence, this volume is surprisingly well-structured, as the author progresses through the lectures in explaining the theory in a coherent way. More importantly, the relevance of this publication is to be seen in the more general enterprise taken up by experimental semantics in recent years, which is to allow experimental work to discriminate among equally plausible theoretical options and to substantiate formal representations by giving cognitive content to logical variables or other abstract objects. In this perspective, this small volume is a welcome contribution for theorists and experimentalists alike, and an additional step toward realising the goal set by experimental semantics.


Since its early days, cognitive linguistics has been evolving from its original place at the crossroads of philosophy, psychology, and semantics to embrace more formal approaches to language. As regards the representation of events in language, the renewed interest in cognitive approaches is witnessed by the vast amount of recent work on causal relations, aspectual modification, and quantification in event semantics (cf. Truswell 2019 for an overview). Following this thread, in this review I will try to highlight the points in Zacks's lectures that answer more specifically some of the most urgent questions addressed by formal linguistics (and in particular formal semantics and syntax) to the topic of event representation. The review is built around a selection of these questions, which may serve as a guide for the reader interested in specific issues; at the same time, by highlighting theoretical issues I will try to show how the work by Jeffrey Zacks and his colleagues provides a possible way of addressing them from the point of view of cognitive psychology.

What is an event?

There is a common knowledge and a shared intuition about what an event is. It is something that happens, and as such has a temporal dimension. It also happens to something or someone, and as such it has participants. Philosophers struggled for decades to frame these intuitions in coherent theories in which linguists have shown a keen interest because of the difficulty inherent in formally defining such an object. Empirical evidence sometimes suggests considering events as logical individuals (Davidson 1967), sometimes treating them as properties of time (Vendler 1967, Verkuyl 1999, a.o.). As Zacks correctly reminds us in his first lecture, however, one of the most pressing questions that arise when looking at the philosophical literature is this: which theories are going to be useful for cognitive science, for linguistics, and for psychology? Research in psychology points to an answer: an event seems to be perceived as a structured representation, that is, as a “segment of time at a given location that is conceived by an observer as having a beginning and an end”. Beginning and end coincide with changes in the most relevant properties of an event. But what are these? In a series of experiments the results of which are commented on in these lectures, Zacks and his colleagues have shown that the properties more likely to be perceived and associated with the individuality of an event are those that help us to build an internal coherence in what we see. This is the basis for the actual hypothesis advanced in Event Segmentation Theory: humans recognise event boundaries by building models (representations of interactions in a spatio-temporal framework) from the events they have experienced, and by anticipating what they are experiencing. Models are then updated regularly in case we make bad predictions. There are two main features of Event Segmentation Theory that I would like to point out.

First, the theory says that our ability to recognise events is rooted in the adaptive comprehension system of human cognition, which is forward-oriented. By making predictions about what they are going to see, humans can filter out everything regular and persistent from the crammed set of data that our sensory organs register in real time, and they can focus on outstanding things, keeping track of changes to the regular behaviour of things.

Second, as it is made clear in the third lecture, the theory makes a claim also about the way in which we learn to do this, which is “in the same way that we learn to comprehend spoken language: by being immersed in it.” What Zacks seems to be suggesting here is that the segmentation of events is acquired through a bootstrapping process akin to that responsible for the prosodic bootstrapping that, according to certain acquisition theories, leads to the segmentation of syntactic units in language acquisition (see Soderstorm et al. 2003). These two assumptions put together inscribe Event Segmentation Theory within the theoretical background shared by some of the most followed linguistic theories, and confirm the psychological stance on language from the point of view of a cognitive scientist.

Segmentation and plurality – how many events?

There is, finally, a third issue. The issue of the identity of events is a delicate one – identity and individuation are a prerequisite for metaphysical properties, such as persistence, and for building abstract logical objects, such as pluralities. As pointed out by Truswell (2019), once we have a theory providing us with a set of mereological relations (as in e.g. Link 1983), from a logical point of view nothing prevents us from generating arbitrary events, by decomposing and recomposing the primitive individuals into arbitrary sets. The issue, then, is “to complement this logic with a characterisation of cognitively relevant events” (Truswell 2019: 96).

Studies investigating the segmentation of events in narratives (expounded in particular in Lecture Four) point out that, in a default situation (i.e., when our attention is not channeled in a specific way), the properties more likely to be perceived as relevant to building an event model mainly concern the participants and the temporal continuity of an event (with spatial contiguity a less salient criterion). What does this mean, more specifically, for linguists concerned with representing events in language? The issue of temporal contiguity and change has been evoked, for instance, in the attempt to motivate the linguistic relevance of distinct aspectual categories according to which verbal predicates may be classified. While certain events (such as playing, sleeping, or laughing) may naturally persist in time without change, and thus are treated as single occurrences, others (such as breaking a glass or writing a letter) necessarily reach a point after which they cease to be describable as the same event. Also, some activities (jumping) can be understood as events persisting in time or as collections of similar events (semelfactive events – like multiple occurrences of making a jump), and in this case, the (absence of) change that makes a plurality of events one single event is probably in the eye of the beholder.

How are events treated by grammar?

Finally, Zacks suggests that, keeping the mapping from mind to language as direct as possible, the structured kind of representation evidenced by psychological studies should be translated by a structured predication. There are different models of predicative structure out in the linguistic world – the baseline being that events, as logical objects, may be conceived of as relations between individuals, which, when happening, leave a temporal trace. Formal linguists then developed models where individuals and time are linked to events by functional elements – individuals by thematic roles and time by the homomorphic function whose value is the running time of the event (see e.g. Krifka 1992 for a formal semantic representation). Evidence from cognitive psychology seems to support the theoretical and formal models developed by linguists, but once again, experimental methods may be seen as pivotal in discriminating among theoretical hypotheses. In particular, Zacks and colleagues argue that their results are best represented formally within the models developed in the framework of situation semantics (Perry & Barwise 1983). But of course several questions are still pending: how many thematic roles do we need, and how should each role be characterised? Other studies grounded in cognitive psychology may help answer these questions (see e.g. the approach developed by Wolff 2007 and subsequent work, based on psychological theories of causation). The work resumed by Zack in this volume, however, has the merit of showing the relevance of such a question for embedding cognition into grammar.


Davidson, D. (1967) The logical form of action sentences. In N. Rescher (ed) “The Logic of Decision and Action”, 81-95. University of Pittsburgh Press.

Krifka, M. (1992) Thematic relations as links between nominal reference and temporal constitution. In I. Sag and A. Szabolcsi (eds) “Lexical matters”. Stanford , CSLI

Link, G. (1983) The logical analysis of plurals and mass terms. A lattice-theoretical approach. In Bäuerle, Schwarze & von Stechow (eds) “Meaning, Use and the Interpretation of Language”, 303-323. Berlin : de Gruyter.

Perry, J. & J. Barwise (1983) “Situations and Attitudes”. MIT Press.


Marta Donazzan is lecturer of Linguistics at the University of Nantes, and researcher at the Laboratoire de Linguistique de Nantes (LLING/ UMR 6310). After receiving her PhD in Linguistics at the University of Paris 7 in 2008, she worked as a lecturer and a researcher in various institutions in France and Germany. Her interests are in formal semantics and the interface between cognition and grammar.

Page Updated: 14-Jul-2021