LINGUIST List 28.4872

Mon Nov 20 2017

Review: Cheyenne; English; Morphology; Pragmatics; Semantics; Syntax; Typology: Murray (2017)

Editor for this issue: Clare Harshey <>

Date: 23-Aug-2017
From: Hanno Beck <>
Subject: The Semantics of Evidentials
E-mail this message to a friend

Discuss this message

Book announced at

AUTHOR: Sarah Elizabeth Murray
TITLE: The Semantics of Evidentials
PUBLISHER: Oxford University Press
YEAR: 2017

REVIEWER: Hanno T Beck, University at Buffalo

REVIEWS EDITOR: Helen Aristar-Dry


Linguists have found that many languages have evidential morphemes (de Haan 2013), morphemes that encode the source of information conveyed by a statement. English does not have evidentials but can express evidentiality through periphrastic means, as for example with “I heard” in “Lee passed the test, I heard.”

Sarah Murray has studied Cheyenne (an Algonquian language with about 2,400 speakers mostly in Montana (Dryer & Haspelmath 2013, Simons & Fennig 2017)), which has three evidential morphemes. Other languages whose grammars include evidentials may have more or fewer morphemes with similar or different meanings. In the case of Cheyenne, Murray explains that one of the evidentials is reportative, meaning roughly “I hear”; a second is direct, meaning roughly “I witnessed”; and a third is inferential, meaning roughly “I gather/infer.” (Of course if used in a question then the morphemes would mean not “I hear” but “you hear” and so on.)

There is ample crosslinguistic variety, but obviously one wants to attempt a unified account of the phenomenon, to give a comprehensive view of how human languages that have evidential morphemes implement the expression of evidentiality.

In particular, semanticists interested in this topic will busy themselves with questions such as: What is the range of meanings expressed by utterances involving evidentials? What sort of semantic theory will make sense of that range of meanings? And can we formalize our observations in a way that accounts for the crosslinguistic data while fitting evidentiality into the broader landscape of formal semantics and pragmatics?

In “The Semantics of Evidentials,” Murray explores these questions, and brings readers a full harvest of answers. Though relying primarily on Cheyenne data, she includes plenty of examples from other languages and, for comparison to a non-evidential language, English. Here is how she describes her goal: “…this book develops an explicit, truth-conditional account of evidentiality and the semantic contributions of all sentence types that does not appeal to separate dimensions of illocutionary meaning.” (p. 5) By excluding explanations that rely on “separate dimensions,” Murray is indicating that she will see how far we can get in spelling out the semantics of evidentials without following the strategy of Potts (2005), who analyzed conventional implicatures with a multi-dimensional approach. As scientists we want our toolkit to be as sparse and simple as we can make it, while still giving sufficiently rich explanations of the data that we face. Can we stay with a single-dimension semantic theory, while encompassing the evidential phenomena found in languages?

The main feature of Murray’s account, as she reminds the reader often, is that on her analysis an evidential sentence provides three types of semantic contribution: at-issue content, not-at-issue content, and an illocutionary relation. In dynamic semantics we ask how an utterance can change a context, and Murray’s investigation leads her to conclude that there are three types of change; these depend partly on the particular evidential morpheme in play.

Here is an English example and a very cursory description of these three components of context change. With “Pat won a prize, I hear” the at-issue content provided is “Pat won a prize” and unlike an assertion, this content is merely being presented but not being actively claimed to be true, nor is it being advocated that we add it to the common ground. The not-at-issue content is the larger clause, “Pat won a prize, I hear,” which normally enters the common ground immediately upon utterance, assuming a sincere report. And the illocutionary relation is a further adjustment to the overall context – in this example, the illocutionary relation is empty because the speaker makes no further commitments. If instead we take “Pat won a prize, I gather/infer,” here the speaker would be undertaking a slightly greater commitment to Pat’s having won a prize; thus there is a non-empty illocutionary relation, which Murray renders as something like “it is at least possible that Pat won a prize.” This is added to the overall context, with the effect of eliminating some worlds from the context set (ruling out the worlds where it is not possible that Pat won a prize).

This brief illustration would seem clearer in context, and if accompanied by a diagram. In her book, Murray provides both of these. Indeed, she is quite liberal in her use of diagrams and the result is that every important example is not only explained but illustrated. To do valuable work in formal semantics, one finds oneself surrounded by very abstract notions – and any time a simplifying diagram can be used, the viewer benefits.

In Chapter One, “Introduction,” Murray introduces evidentials and outlines her plan for the book.

In Chapter Two, “A semantic classification of evidentials,” Murray conducts a survey of evidentials as used in four languages: Cheyenne, Cuzco Quechua, St’át’imcets, and German. Here we can see into the workshop of a linguist -- what characteristics she comes up with as the crucial diagnostics and behaviors associated with evidentiality, and how these generalize crosslinguistically. These findings become criteria or measuring sticks that can serve in the evaluation of alternative treatments of the data.

In Chapter Three, “Evidentials and varieties of update,” Murray develops arguments for what a theory must do to account for the range of phenomena reviewed in the previous chapter. This is where she provides a full explanation of her proposed system of three semantic contributions for a sentence; how evidentials play a role in interrogatives as well as declaratives; and how her model handles crosslinguistic variation.

What exactly does an evidential do? When you utter an evidential sentence, what are you accomplishing? This is what Murray explores in this chapter -- the effects achieved by evidentials and how to regard these. If an analysis of the semantics of some linguistic phenomenon, in this case evidential morphemes, is to be regarded as more than just one person’s opinion, the analysis should be formalizable. Murray gets right down to brass tacks and shows in the next two chapters precisely how this can work, and thus how her viewpoint can fit usefully with existing theories of formal semantics.

In Chapters Four and Five, “Declarative sentences” and “Interrogative sentences,” Murray presents the formal implementation of the account developed in the previous chapters. Her framework of choice here is Update with Modal Centering (see especially Bittner 2011 for more on this family of frameworks), but she points out that other frameworks could also be up to the task. There is nothing in Murray’s findings that forces her to pick Update with Modal Centering; if you are more familiar with an alternative, you can imagine that as you read. But a framework will, at a minimum, need some mechanism for representing the effects of asserted content; content that is presented but not asserted; content that is not-at-issue; and questions. I am curious to see how Murray’s approach would translate into a suitable version of the context framework envisioned in Farkas & Bruce (2010), but there is no reason to anticipate problems with that.

Why does Murray spend a chapter focusing on evidentials in interrogative sentences? It’s a self-imposed test. If Murray’s account is roughly right, then it should apply well to sentences with other kinds of mood beyond the declarative. Ideally, an account of evidential interrogatives will differ from that of evidential declaratives only in the same ways that non-evidential interrogatives differ from non-evidential declaratives. Murray does uncover some mysterious behaviors here in Cheyenne, looking at polar questions and content questions. The content questions with evidentials show a peculiar ambiguity. This is discussed in depth, and we are presented some options for how the different readings can be formalized; still, it is not clear how we predict the acceptable interpretations over the unacceptable ones. This might be a fruitful area for future research. Some constraint or principle may be awaiting discovery, and implementation within the Update with Modal Centering framework or one of its competitors, that will yield further insight on the puzzle.

In Chapter Six, “Conclusion,” Murray summarizes very briefly what she has attempted in this book, again taking care to keep the reader oriented.

After the conclusion, but before the references and index, there are two appendices. Appendix A, “Definitions and worked examples,” gives us some further examples and also provides a synopsis of Update with Modal Centering. I would have liked to see an example constructed completely, starting with lexical items and building up an evidential sentence from there, to demonstrate the compositionality of this system of context-update contributions.

Appendix B, “Semantic contributions by phenomenon,” presents charts showing Murray’s proposed three context-change effects for a variety of evidential and non-evidential sentences, giving both the semantic tasks performed and how those are implemented within her specific adaptation of Update with Modal Centering.


This is a good book, and more than that, a successful book. That is, the author clearly sets out a specific mission and then proceeds to fulfill the mission. As a reader, I never felt lost or at a dead end. Murray takes great care to keep the reader oriented, to remind us of the big picture while we are looking at particular data points.

The book’s material is very clearly presented. Again and again, Murray hearkens back to one or another of her main points. In fact, although not primarily intended for the classroom, it would be easy to use this book as part of an advanced semantics/pragmatics course. It illustrates what a semanticist does and how the work proceeds; brings the reader up to date on approaches to evidentials; and demonstrates an application of the Update with Modal Centering framework.

One also should commend Oxford University Press, which is energetically publishing books in its “Oxford Studies in Semantics and Pragmatics” series. The presentation is attractive, and I found almost no typographical errors.

A minor confusion that could have been avoided: in this book, numbered examples, tables, and figures all use the same numbering style. It is needlessly messy when there exists an example 5.2 illustrated with table 5.1, while table 5.3 explains example 5.9, and figure 5.2 elaborates example 5.7. A distinct numbering style for each series (letters, Roman numerals, etc.) would be preferable to my eye.

In some places Murray points out how introducing updates in a different sequence would result in different interpretations. What, then, are the principles that determine an update sequence? A sentence can contribute multiple updates to the context, so we need rules to guide the order in which these are processed. If there are generalizations, what are they, and are they fixed or typologically variable across languages?

I also wonder if the treatment can be made more sensitive to the particulars of context. For example, Murray’s description of the illocutionary relation contributed by the inferential evidential – that the at-issue proposition is at least possible – is certainly a good first approximation, but in the reality of conversations there is a very wide continuum of interpretations available for that element. Some inferences are quite tentative, while others are very firm. If the meaning contributed by the evidential is a point along some sort of continuum, is it the theory’s obligation to reflect that, or do we simply have a vague morpheme in the same way that the adjective “tall” is vague?

And more generally, if we will be relying on a notion of three types of context update, I want to learn a bit more about that third one, the illocutionary relation. What is the set of possible illocutionary relations, is that set constrained in interesting ways that tell us things about human language and the mind, etc., because we want it to be a source of insight, not a collection of ad hoc processes for which the theory had no other place. Addressing this in depth would have taken Murray too far from her core topic in this book, but I do believe that these bolts should be tightened up at some point, if illocutionary relations are to play a role in analyzing other linguistic phenomena.

A positive experience is supposed to leave you wanting more. This is the situation with “The Semantics of Evidentials.” Indeed, as it is a fairly short book (172 pages), one wishes for a few extra pages dedicated to discussion about where the author believes we should go next. Which languages are most in need of exploration, which phenomena are still not well understood, what issues lie at the boundary between evidentiality and modality, what enhancements to Update with Modal Centering are worth considering, etc.

What should semanticists say about the bundle of phenomena known as evidentiality, particularly as embodied in evidential morphemes? This book provides a promising and attention-worthy answer, delivered with clarity.


Bittner, Maria. (2011) Time and modality without tenses or modals. In Renate Musan and Monika Rathert (eds.), Tense Across Languages, pp. 147-188. Tuebingen: deGruyter.

Dryer, Matthew S. & Haspelmath, Martin (eds.) 2013. The World Atlas of Language Structures Online. Leipzig: Max Planck Institute for Evolutionary Anthropology.
(Available online at, accessed on 2017-08-17.)

Farkas, Donka, and Bruce, Kim. (2010) On reacting to assertions and polar questions. Journal of Semantics 27(1), 81-118.

de Haan, Ferdinand. (2013) Semantic Distinctions of Evidentiality. In: Dryer, Matthew S. & Haspelmath, Martin (eds.) The World Atlas of Language Structures Online. Leipzig: Max Planck Institute for Evolutionary Anthropology. (Available online at, accessed on 2017-08-23.)

Potts, Christopher. (2005) The logic of conventional implicatures. Oxford Studies in Theoretical Linguistics. Oxford: Oxford University Press.

Simons, Gary F. and Charles D. Fennig (eds.) (2017) Ethnologue: Languages of the World, Twentieth edition. Dallas, Texas: SIL International. Online version:


Hanno Beck is a PhD student in the Department of Linguistics, The University at Buffalo (State University of New York). He is currently teaching linguistics and German, while pursuing research on “subjective” language use, as well as the semantics and pragmatics of the tough-construction.

Page Updated: 20-Nov-2017