Susan Stebbing on Logical Positivism and Communication

1. Introduction

It is often taken for granted that A. J. Ayer is responsible for introducing the main ideas of the logical positivist movement into the British philosophical context. However, in the years before Language, Truth and Logic was published in 1936, Moritz Schlick and Rudolf Carnap gave lectures in London, and Carl Hempel and Schlick had a back-and-forth in Analysis about the protocol-sentence debate.¹ Susan Stebbing is the throughline here. She met Schlick in 1930 at the Seventh International Congress of Philosophy in Oxford (Chapman 2013: 81);² invited both him and Carnap to lecture in London; reviewed Language, Truth and Logic and several of Carnap’s books; and co-founded Analysis.³

In this paper, I focus on a series of Stebbing’s articles and reviews in the early to mid-1930s on logical positivism. I bring to the foreground one aspect that has gone mostly unremarked on in the growing body of work on Stebbing,⁴ namely, her focus on whether the principle of verifiability can ground a plausible account of communication. Stebbing thinks not, and I show how her criticisms are related to other parts of her engagement with logical positivism, including her rejection of methodological solipsism and her dissatisfaction with the logical positivist treatment of statements about other minds and the past. I also argue that Stebbing’s criticisms provide a bridge to later philosophers like P. F. Strawson and J. L. Austin who characterize their ordinary language approach to philosophical analysis by contrasting it with logical positivism.

Attending to Stebbing’s work sharpens our understanding of the early British reception of logical positivism prior to Language, Truth and Logic and draws attention to early criticisms of the principle of verifiability more interesting than the well-worn objection that it fails to be meaningful by its own criteria.⁵ As Michael Beaney (2016) also notes, given Stebbing’s prominence and that her work was published in high-profile venues like Proceedings of the Aristotelian Society and Mind, she surely had a role in shaping the reception of logical positivism. Finally, following Richardson (2017), I think foregrounding Stebbing’s work paints a fuller picture of how logical positivism came to be part of analytic philosophy despite having different concerns than many of the early analytic British philosophers.

2. Knowing, Understanding, and Analyzing

In the 1930s, Stebbing wrote a series of articles devoted to distinguishing different types of philosophical analysis and their strengths and weaknesses.⁶ It is in this context that she considers logical positivism. Stebbing is broadly sympathetic to the then-new philosophical movement given the central role its proponents assign to analysis in philosophical practice. To Stebbing, the logical positivist approach to philosophical analysis seemed similar to that inspired by Moore, yet potentially distinct in their understanding of ‘analysis’ and ‘facts’ (1933a: 6). Her 1933 British Academy lecture, “Logical Positivism and Analysis” is concerned, then, with understanding Wittgenstein’s conception of analysis “in relation to the philosophical practice of Moore” (5). However, because “the Tractatus Logico-Philosophicus, is excessively condensed and, no doubt deliberately, oracular,” Stebbing relies on the work of Schlick and Carnap, among others, to illuminate Wittgenstein’s views (4).⁷

Before turning to Stebbing’s reconstruction of the logical positivist conception of analysis and its differences from Moorean analysis, it is worth noting that historical work in the last three decades has disentangled the views of the philosophers then-referred-to as logical positivists, revealing substantive philosophical disagreements.⁸ So, we might worry Stebbing has made a mistake in taking Schlick, Carnap, and Wittgenstein to offer a unified conception of philosophical analysis that can be fruitfully contrasted with Moore’s. Perhaps! And the diversity of views among the logical positivists would need sorting through were my main interest in how Stebbing’s criticisms hit or miss their mark. Insofar as I am interested in understanding how logical positivism was first received in early analytic British philosophy, I simply note Stebbing lumps them together, though not without reason. In a footnote, she draws attention to an “interesting statement, from a manifesto published by ‘the circle’ in 1929” (4 n.1). The statement from the preface to the manifesto says:

This circle has no rigid organization; it consists of people of an equal and basic scientific attitude; each individual endeavours to fit in, each puts common ties in the foreground, none wishes to disturb the links through idiosyncrasies. In many cases one can deputise for another, the work of one can be carried on by another. (Neurath et al. 1929/1973: 299)⁹

Stebbing, it seems, took this remark seriously, and so thought Carnap’s and Schlick’s less cryptic work could clarify Wittgenstein’s views on analysis.¹⁰

Stebbing begins the substance of her lecture by characterizing three commitments of Moore’s approach to analysis. These three commitments, according to Stebbing, clash with the central role verifiability plays in the logical positivist conception of analysis, with the major difference being that Moore seeks to preserve a tripartite distinction between knowing a proposition, understanding it, and giving an analysis of it. The first commitment of Moore’s approach is that “at various moments in our lives we are in a position to assert with regard to a certain proposition that we know this proposition to be true” (Stebbing 1933a: 6–7). Here, Stebbing highlights Moore’s contention that philosophical analysis is not in the business of justifying our knowledge of certain propositions or showing them to be true. The second commitment is that “with regard to many such propositions there are expressions in ordinary usage which unambiguously express these propositions which we know to be true. A proposition is unambiguously expressed when what is said is understood” (7). For some of those propositions we know to be true, there is an ordinary language equivalent expressing it that can be understood by competent speakers, though with varying degrees of clarity. Finally, the third commitment is that “to understand an expression is not equivalent to being able to give a correct analysis of its meaning” (7).¹¹ This last point sets up “the chief task of philosophy” for Moore, which, according to Stebbing, “is to discover the correct analysis of expressions which every one would agree are sometimes used to say what is true” (9).

As Beaney (2016) also points out, one of Stebbing’s main complaints is that the logical positivists collapse the tripartite distinction between knowing, understanding, and analyzing central to Moore’s account. For example, Stebbing draws attention to Schlick’s claim that

In order to understand a proposition we must be able exactly to indicate those particular circumstances that would make it true and those other particular circumstances that would make it false. ‘Circumstances’ means facts of experience; and so experience decides about the truth or falsity of propositions . . . (Schlick 1932/1938b: 142–43)

Understanding a proposition involves knowing what experiences would lead us to judge it as true or false. Elsewhere, Schlick connects understanding to knowledge and meaning. He says,

But how can we be quite sure that we really know and understand what we mean when we make an assertion? What is the ultimate criterion of its sense? The answer is this: We know the meaning of a proposition when we are able to indicate exactly the circumstances under which it would be true (or, what amounts to the same, the circumstances which would make it false). The description of these circumstances is absolutely the only way in which the meaning of a sentence can be made clear. (1932/1938a: 126–27)

Schlick, then, identifies understanding and making clear the meaning of a proposition; both consist in giving an account of how we would come to know it as true or false, that is, in specifying its verification conditions.

Stebbing sees Schlick as running together what Moore is at pains to keep separate. Schlick claims that “before the sciences can discover the truth or falsity of a proposition they have to get at the meaning first”; as such, “they cannot go on with the pursuit of truth before the pursuit of meaning has become successful” (130). Stebbing criticizes this view. She claims,

Understanding more or less unclearly what we say, we nevertheless may know that what we say is true. We then inquire what must be the case if what we have said is true. In this way we may come to see more clearly what it is we were knowing. . . . [B]ut it is a muddle to suppose that this clarification is a pursuit of meaning. The word ‘meaning’ is too ambiguous, unclear, and vague, to be helpful in this connexion. (1933a: 33–34)

For Stebbing, we can know when a proposition is true and understand the sense of the proposition with varying degrees of clarity without being able to analyze that proposition in terms of its verification conditions. Moreover, even if an analysis of the verification conditions may help us “come to see more clearly what it is we were knowing,” it does not constitute an analysis of meaning.¹²

Thus, Stebbing understands the central disagreement between the Moorean and logical positivist pictures of analysis as follows. For Schlick, an analysis of a proposition in terms of its verification conditions clarifies its meaning. In grasping the verification conditions, we then understand the proposition. An analysis of a proposition’s meaning comes before we can know the proposition because it is only once we grasp its meaning that we can then go on to determine its truth or falsity by seeing if its verification conditions attain in experience. Stebbing suggests, though, that the propositions submitted to analysis, as expressed unambiguously in ordinary language, are already known to be true. In a different paper, Stebbing maintains analysis is not concerned with “find[ing] reasons for our beliefs” (1932–1933: 70). Instead, our knowledge “that there is a table in this room, or that I am now sitting at this table, or that putting my hand in the flame was unpleasant, must afford a starting point” for analysis (70).¹³ The results of analyzing these propositions is clarification of

(1) what exactly we are believing when we believe that there is a table in the room, that it was here three hours ago, and so on; (ii) how our various beliefs are inter-related; (iii) how our inconsistent beliefs may be adjusted, and which should be rejected. (70)

If the propositions that form the starting point of analysis are known to be true and are unambiguously expressed, then how can they be understood with more or less clarity? Stebbing claims “an unambiguous expression is not equivalent to a perfectly clear statement” (1933a: 7 n.3). Elsewhere, she says, “A word is understood when it is recognized as a sign signifying a referend” (1930/1942: 13–14). For Stebbing, “a word is ambiguous when it stands for different referends on different occasions” (21). An unambiguous expression that is understood, then, seems to be one whose immediate reference is recognized. When I say ‘There is a table in this room,’ the table, room, and the spatial relationship to which I refer are recognized (or are recognizable). So, the expression is unambiguous. I can also know that the proposition expressed by the sentence is true. However, following Moore, Stebbing claims this does not mean I thereby know what an analysis of the proposition consists in. For example, I might not know what other facts it commits me to or how it is related to sentences expressing beliefs like ‘This is a table,’ ‘I see a table,’ and ‘This table is made of wood.’¹⁴ By showing what is involved in knowing ‘There is a table in this room,’ analysis clarifies exactly “what must be the case if what we have said is true” (1933a: 34).¹⁵ For Stebbing, this process does not result in verification conditions that give the meaning of the analyzed proposition; after all, its meaning is already understood. Instead, analysis clarifies what is involved in knowing the proposition in part through relating it to other beliefs we have, especially about things that are or can be presented in experience.

So far, I have just stated Stebbing’s criticism that the logical positivists run together understanding, knowing, and analyzing a proposition.¹⁶ To fully understand it requires we examine her reconstruction of the logical positivist use of the principle of verifiability, as well as how she sees it underwriting a particular account of communication. Ultimately, she thinks their account has significant shortcomings due to its commitment to methodological solipsism (see §3) and its analysis of statements about other minds and the past (see §4).

3. Communication in Logical Positivism: Verifiability and Solipsism

Stebbing’s criticism that the logical positivists mistakenly conflate knowing, understanding, and analyzing a proposition forms the background against which she develops her criticisms about the implications of their view for communication. A few words about Stebbing’s focus on communication are in order since it might seem that the logical positivists do not aim to offer any such account, but are rather, to put in a way in which Carnap might, rationally reconstructing the logic of science.¹⁷ In short, Stebbing thinks that the use of “the ‘new logic’ of Russell and Wittgenstein” by the logical positivists “has turned their attention to symbolism, i.e., to problems of the structure of language and the possibility of constructing deductive systems” (1933a: 15). Moreover, for Stebbing, “Language is to be understood as any means of communication; it is not directly presentative or pictorial, for language does not present what it is used to say, but communicates it” (16). Stebbing thinks that insofar as the logical positivists are interested in “symbolism,” their approach to analysis has implications for the use of language to communicate, even if they intend mainly to explicate the logical relationships between propositions.

Indeed, Schlick’s 1932 London lectures open with references to communication and its importance to human activities. Though Stebbing did not attend the lectures for health reasons, she did receive a verbatim report from her student Margaret MacDonald, from which she quotes. Schlick says in the opening to his lectures, “There would be no cooperation between human beings if man could not exchange ideas with his fellow men; there would be no arts, no science, if knowledge could not be handed down from one generation to the next” (1932/1938c: 152). For our purposes, it is important to emphasize that Schlick thinks form or structure is always what is communicated. Form, the type of order that logic is concerned with (158), is something that can be shared between two people, as well as between language and the world. However, what Schlick calls content, that which we are acquainted with in our sensed experience of the world, is incommunicable, since it “is furnished by the individual himself, derived from his own experience” (164). As Schlick puts it, “The difference . . . between form and content is, roughly speaking, the difference between that which can be expressed and that which cannot be expressed” (159).

As Stebbing (1933a: 16) hints, Schlick seems to accept something like Bertrand Russell’s distinction between knowledge by description and knowledge by acquaintance, though with some hesitance to call the latter ‘knowledge.’ Schlick says, “all genuine knowledge is Expression” (1932/1938c: 187). This separation between form/structure and content, with the latter inexpressible and the former expressible, hence communicable, is a theme throughout logical positivist work at the time. For example, in their overview of logical positivism in the The Journal of Philosophy in 1931, Albert Blumberg and Herbert Feigl mention Russell’s distinction, then say, “knowledge or the communicable expresses the formal structure but not the content of experience. For the immediately given is private, non-communicable” thereby not counting as knowledge proper (1931: 285).¹⁸ To echo Stebbing, language does not present content, since that is private to each individual, but instead communicates structure, which is capable of being shared.¹⁹ Thus, she thinks that

for Logical Positivism, the problem of knowledge resolves itself into the problem how language can be used to communicate. This is just the question: How do sentences mean? To which, from this point of view, the reply must be that sentences mean by conveying structure. (1933a: 18)

What, then, is the connection between structure and content? Here, finally, enters “Wittgenstein’s principle of verifiability: the meaning of a proposition is the method of its verification” (18). After stating the principle, Stebbing asks, “How could a proposition be verified if what is communicated is never content (Inhalt) but only structure? Clearly, verification depends on the presence of content,” that is, sensory experience (18). And while Schlick offers an extended discussion of the verifiability principle, Stebbing turns to Carnap since, “It is to Carnap that . . . the fullest development of this view is due” (19).

In particular, she considers Carnap’s views about the relationship between the universal, physical language and protocol-languages put forward in “Die Physikalische Sprache als Universalsprache der Wissenschaft,” translated into English in 1934 by Max Black as “The Unity of Science” (Carnap 1934a). On Stebbing’s gloss of a protocol-language, it is “a language which directly reports my own direct experience” (1933a: 19). Since such languages are grounded in our private experiences, or in content, Stebbing takes them to not overlap. She says, “there are as many protocol-languages as there are experients. This, it seems to me, follows directly from Wittgenstein’s interpretation of the principle of verifiability” (20). On Stebbing’s understanding, an analysis of a sentence in terms of verification conditions bottoms out in my protocol-language, which, being based in an individual’s private experience, cannot be shared. This analysis gives the meaning of any proposition, and, following Schlick, understanding the meaning allows us to determine if the proposition is true or false. So, Stebbing says of this picture,

The meaning (Sinn) of the proposition I express by saying ‘This is a table’ is just how it would be verified, i.e. how it would be found true and how it would be found false. This verification must lie within my own experience. (19)

Thus, what verifies and gives the meaning of the statement ‘This is a table’ is a statement expressed in my protocol-language about experiences I am currently having or could have.

What are the implications of this view for the possibility of communication? How, according to the logical positivists, do I understand what others mean? Consider that if someone else were to report being thirsty, my report of their thirst is verified by some statement about my experience, made in my own protocol-language, and not by their experience (21). I would not, then, seem to understand someone’s report of their thirst. Yet, we do manage to understand one another. On Stebbing’s reconstruction of Carnap, this is accomplished by shifting to the formal mode of speaking about words away from the material mode of speaking about the world. To solve the issue of communication, we must take a protocol-language “as expressing structure, not as expressing content, as containing words, not as describing states of affairs” (21). We end up communicating, and thus understanding each other, by transforming our private protocol-languages into the shared physical language. This shows how “your protocols and my protocols are all sub-languages (Teilsprache) of the universal, physicalistic language, although these various sub-languages contain nothing in common” (22). For the logical positivists, on Stebbing’s reading, communication is possible because there exists a relationship between the shared physical language and individual protocol-languages that allow us to express structure even if, ultimately, verification bottoms out in private experience.

Stebbing gives the following example: “a protocol proposition ‘Brown now’ seen by me could be deduced by you from a sufficiently definite description of my body” in the shared physical language (22). While a proposition in the physical language about my body might seem different from what my protocol sentence says about my own private experience, so long as the statements are mutually deducible, that is, can be translated into one another (22), they share structure. In Stebbing’s words,

[M]y protocol-language, which never overlaps with your protocol-languages, can nevertheless be understood by you, since my protocol can be transformed into the language of physics. It is true that a proposition p, in my protocol, which only I can understand, will not seem the same to me as the proposition p´ into which it is transformed in the physicalistic language. This is because p is associated with my own protocol, and p´ with the physicalistic language. Nevertheless, if p and p´ are mutually deducible, they express the same, and what they express is structure. (22–23)

Stebbing is skeptical about this account for a number of reasons, one of which is that it seems to suggest that there is something right about solipsism.²⁰ She recognizes that Carnap’s solipsism is methodological, relating to the logical construction of the objects of experience, rather than metaphysical. However, she does not think this avoids certain problems. Noting that throughout her lectures she has intentionally used ‘we’, she says “I have the best of grounds for denying solipsism namely, that I know it to be false. You, who are listening to me, and enable me to speak in the plural, also know it to be false” (27). Here, she cites Moore (1925) in a footnote, connecting her denial back to the first commitment of Moore’s picture of analysis reconstructed in §2: there are some things we know to be true. One of those things I know to be true is that I share a world with others. Stebbing grants that perhaps something like Carnap’s methodological solipsism and its corresponding “theoretical, abstract system may be adequate to describe what I suppose to be happening to other minds, or to bodies” (1933a: 26); it might also serve as a basis to “try to order my own experience in a system based upon . . . directly given experience, and constructed by means of logical operations” (26–27). Yet, however useful it might be for the purposes of building an abstract system with as few assumptions as possible, Stebbing thinks it fails given the implications it has regarding the possibility of communication.

Stebbing makes this point clearly in response to R. B. Braithwaite’s criticism of her appeal to common sense truths to reject methodological solipsism. Braithwaite thinks we can understand methodological solipsism not as “asserting propositions incompatible with the ‘Common Sense view of the world’,” but as “propound[ing] analyses of these propositions” (1933: 14). We cannot reject methodological solipsism by simply appealing to what we know to be true since it makes no statements about the world. More importantly, it can be seen as giving an analysis of common-sense truths. Stebbing grants this criticism. However, she maintains that the logical positivist analysis of the meaning of statements in terms of verification conditions fails at “making clear what is involved in ‘understanding’ and in ‘communication’” (1934a: 28).

For one, according to Stebbing, even if certain facts expressible in the physical language about my body in relation to the table in front of me are conveyed in the statement, ‘This is a table,’ such facts about my body are not communicated. The distinction between communicating and conveying is fleshed out more fully in §4. The second point to note is that those conveyed facts about my body do not constitute the meaning of the original sentence. Put differently, facts about my body expressed in the physical language are not what we understand when someone states, ‘This is a table,’ even if facts about my body might figure into an analysis of its verification conditions.²¹ Indeed, to recall the distinctions made in §2, Stebbing thinks I can understand the sentence and know it to be true without giving an analysis of it in terms of my experiences. Third, Stebbing maintains “This table is not an experience of mine. Hence, in saying ‘I perceive this table’, I am not saying ‘I perceive an experience of mine’” (1933a: 28). I experience the table, which is neither an inferred entity nor a logical construction. In the next section, we see that Stebbing maintains that some common reference is needed to explain the possibility of communication. Her claim here is that the world provides the common reference necessary for communication, not a shared physical language expressing structure from which protocol-languages are deducible.

At this point in “Logical Positivism and Analysis,” Stebbing distinguishes four types of analysis, arguing that logical positivism is not sensitive to differences between the types, especially what she calls directional analysis and its commitment to basic facts.²² Much has been written on Stebbing’s distinctions, their strengths and weaknesses, and her emphasis on directional analysis; I refer the reader to that work.²³ In what follows, I look at Stebbing’s account of communication, its implicit denial of solipsism, and her claim that it handles statements about other minds and the past better than the principle of verifiability.

4. Stebbing’s Account of Communication: Other Minds and Historical Statements

In a 1934 paper titled “Communication and Verification,” Stebbing takes up the question about the possibility of communication in more detail.²⁴ The paper deals again with Schlick and Carnap and adds C. I. Lewis and Ayer into the mix. Here, we see the different threads of Stebbing’s criticisms come together, especially as illustrated by her treatment of statements about other minds and the past.

Stebbing opens her paper by claiming, in a way that recalls Schlick (1932/1938c: 152), “The possibility of both commonsense and scientific knowledge presupposes that different people communicate” (Stebbing 1934b: 159). A language is what allows us to communicate. Stebbing defines a language as follows: “‘any set of objects arranged in a determinate way, in accordance with a rule, so that the arrangement of a determinate set of objects in a determinate way represents another set of objects arranged in another determinate way’” (159).²⁵ She notes that since a language is a set of objects arranged in accordance with a rule, “there is no language apart from a user” (159). This user intends to use some arrangement of objects, “‘a sentence’,” put together in accordance with a rule, to represent some other arrangement of objects “‘a fact’”; call it “F₁” (159). This brings us to a key distinction. According to Stebbing, a speaker who uses a sentence to represent some fact, F₁, intends to represent, and so communicate something about F₁. However, in doing so, the speaker also conveys information about their own awareness of F₁. Stebbing says,

This distinction is important because, unless F₁ is a fact about the speaker (e.g., I am tired, I see a ship), what the speaker primarily communicates is not a fact about himself, although he always conveys a fact about himself; when he primarily communicates a fact about himself he also conveys still another fact about himself. (160)²⁶

With this distinction in mind, Stebbing says,

To answer the question of how communication is possible is to answer the question how sentences can be used by a speaker A, so as to refer a hearer B, to that which A intended to refer B. Unless there is such common reference communication has failed. When a successful communication is made by A to B, then B understands what A intended him to understand. (160)²⁷

In taking up this problem, Stebbing highlights four different elements of communication. She thinks, “Communication requires at least (1) a communicator, (2) something communicated, (3) a language, (4) a recipient of the communication” (160). While she grants that communication with oneself is possible, it must be the case that the person plays two different roles. I read this as suggesting that intrapersonal communication is parasitic on interpersonal communication.²⁸ If this is right, then her account of communication functions as an implicit denial of solipsism, much like she intended the use of the plural first-person in her lecture on logical positivism.

Putting this aside, her focus on intention and common reference is central to her account of how communication is possible. When a speaker communicates to a hearer a sentence referencing the chair in front of them, the speaker intends to refer to the chair, not their experiences or awareness of the chair. That said, in communicating a fact about the chair in front of them, the speaker does convey further facts about their awareness, namely that they are having chair-like experiences (or sensing a chair). Stebbing worries that the logical positivists run together what is communicated and what is conveyed in their analysis of statements in terms of verification conditions. On her understanding, since the meaning of a statement is whatever verifies it and verification bottoms out in some experience of mine, then the fact communicated by any given sentence is a fact about the speaker’s experience of the chair as expressed in the shared physical language. Notice that the fact communicated is something about structure—the speaker’s body relative to the chair—rather than the content of their experience, since content cannot be shared in common. However, statements about the speaker’s body relative to the chair—structure—can be shared. With this understanding of the logical positivist position’s implications for communication, Stebbing says, “I must admit that the proposition, ‘Mr. A’s body is chair-seeing’ does not seem to me to elucidate the proposition ‘Mr. A is seeing a chair’” (165).²⁹ But, she continues, “It is not difficult to see, however, why the former must be substituted for the latter once the principle that the meaning of a proposition is the method of its verification has been accepted” (165).³⁰

However, the speaker intends to talk about the chair—an object in the world—and the hearer understands this intention given the unambiguous sentence they choose to employ—‘This is a chair’—rather than a different, unambiguous sentence that would indicate they intend to refer to their own experience—‘I am seeing a chair.’ Stebbing aims “to maintain . . . that there is a fundamental difference between propositions which would ordinarily be said to be propositions about myself . . . and propositions made by me, but not about myself” (168). Maintaining this distinction, on its own, does not settle the question of how we might analyze the sentence, but we do not need such an analysis either to know the proposition or to understand it as referencing a particular chair. Here is an important connection to Stebbing’s worries that the logical positivist conception of analysis collapses knowing, understanding, and analyzing a sentence with its emphasis on the principle of verifiability as a theory of meaning.

Stebbing thinks the problems arising from this collapse are particularly apparent for statements about other minds and historical statements, both of which she sees as central to communication (170). First, consider statements about other minds. For Stebbing, I cannot experience your toothache—she accepts the privacy of first-personal experience—but your toothache is the sort of thing which I could experience. Whereas we can all directly refer “to an actually occurring ache” when we have one, when I speak about your toothache it is my intention to “refer indirectly” to your experience, not my own (168). Moreover, a statement about your toothache is “significant for me because I could directly refer to an actually occurring ache” were I to have one (168). That is, for Stebbing, the meaningfulness of a statement about your toothache—and, analogously, statements about a toothache I might experience in the future (168)—is not found in the experience that would verify it, for example, my experiencing your bodily behavior accompanied by you stating ‘I have a toothache.’ Instead, the meaningfulness “depends upon my ability to recognize an actually occurring ache when it occurs, not upon . . . directly referring to an actually occurring ache” (169). And while I cannot directly verify the statement since I cannot have your experience, “I can understand it as referring to an experience which I do not have” because the experience is of the sort I could recognize were I to have it even though I am not currently (169).

Now, consider Stebbing’s treatment of historical statements.³¹ In her lecture, Stebbing thinks the analysis of historical statements by the logical positivists goes something like this:

Every proposition which apparently asserts a fact with regard to the past, e.g. Queen Anne died in 1714, asserts a number of hypothetical facts with regard to my own future experience, e.g. that if I consult such and such records I shall find that certain statements have been made to the effect that a certain event happened at a certain date. (1933a: 27)

However, even if this might be an account of how I come to know the truth of ‘Queen Anne died in 1714,’ it does not give an account of our understanding of the meaning of the statement. Its meaning, which is the object of understanding, is not identical to whatever would allow me to determine its truth or falsity. In line with the earlier tripartite distinction, Stebbing maintains that questions about knowing and understanding “must be sharply distinguished” (1934b: 170–71); I can understand statements about the past perfectly well without knowing the particular evidence that would verify it. This is because statements about the past are about things that we could experience even if we are not currently experiencing them and never can. Statements about Queen Anne’s death are “about the sort of thing which I could experience now, although, as a matter of fact, I am not experiencing it now” (171). Again, the claim about Queen Anne’s death is not intended to be about my experience, though it is through experience that we would come to know the statement and verify it to be true or false.³² Instead, just like statements about other minds, the statement is

about the sort of thing I could experience, even though it should be the case I never can actually be having the experience itself. It is in this sense only, I think, that all expressions which I can understand must ‘ultimately refer to my experience’. (171)³³

What does Stebbing have in mind here? She appeals to Moore to help clarify (170). Moore says “that what is of the same kind as an object of experience is just what can (in one sense) be experienced, although, as a matter of fact, it never can (in another sense) be experienced” (1902–1903: 93). Here, the first sense of ‘can be experienced’ seems to mean ‘the sort of thing capable of being experienced by beings like us’; the second sense seems to mean ‘the sort of thing we are practically able to experience.’ Historical statements about Queen Anne’s death are about events we can experience in the first sense, but not the second. We are capable of experiencing the deaths of others who presently exist and Queen Anne’s death is of the same sort as deaths in the present. But we are not practically able to experience Queen Anne’s death given that it happened in the past. Similarly, for statements about other minds. Your statement ‘I have a toothache’ refers to a type of experience I am capable of having and recognizing—a pain in my tooth—though I cannot experience a pain in your tooth. Nonetheless, given that I am capable of referring directly to an experience which is of the same sort as yours, namely, a toothache, I am also able to understand statements you make about your toothache and so refer indirectly to it.³⁴

It is, for lack of a better term, the same-sortness of experience that supplies the common reference needed for communication, that is, for understanding the object of someone’s intended reference. And that same-sortness is supplied by the fact that, when we communicate, our choice of expression makes clear when we intend to refer to the world and when we intend to refer to our experience. For Stebbing, questions about what is communicated by statements about other minds and about the past concern understanding someone’s intended reference, not how we would come to know them as true or false or how we would analyze them. And understanding what someone is referring to is supplied by the same-sortness of experience. On Stebbing’s view, verification conditions might tell us something about how to establish the truth or falsity of statements about the past and other minds, but they do not provide meaning. Thus, at least part of Stebbing’s criticism seems to be that the logical positivists understand ‘verification’ in a non-standard way when they put forward the principle of verifiability as a theory of meaning. MacDonald, whose contemporaneous work I compare to Stebbing’s in the footnotes in this section, puts the point clearly: “verification is usually employed in science and elsewhere, not to establish the meaning of propositions, but to prove them true” (1934: 144).³⁵

So, recalling the tripartite distinction canvassed in §2, Stebbing recommends we keep separate questions about understanding someone’s intended meaning in using a particular sentence, knowing whether what is said is true, and analyzing what is said. According to her, “We certainly cannot have knowledge of that which we do not understand, but we can understand that which we neither know nor believe” (1934b: 171). Moreover, while “The ultimate verification of every proposition must be such that it could be expressed in a proposition about my experience,” Stebbing claims “it does not follow from this that every proposition is in fact about my experience” (173). If we collapse understanding, knowing, and analyzing, Stebbing thinks we will fail to give an adequate account of communication and fail to do justice to the fact that we understand the meaning of statements about other minds and the past.

With Stebbing’s criticisms of logical positivist analysis and its implications for communication, as well as her alternative picture of analysis and communication in place, I turn now to consider her relationship to ordinary language philosophy.

5. Stebbing and Ordinary Language Philosophy

One striking thing about the work I have used to reconstruct Stebbing’s criticisms is that, as Richardson says, “The characteristic concerns coming from the methodology of the exact sciences animating the work of Schlick and Carnap are almost entirely absent” (2017: 151). The new logic provided, in the words of the Vienna Circle’s manifesto, “a symbolism freed from the slag of historical languages” (Neurath et al. 1929/1973: 306). Its use in analysis was not to explain the possibility of communication or provide an analysis of what we understand when we talk with one another. The logical positivists were also not out to capture how ‘verification’ is ordinarily employed within the sciences and elsewhere when they put forward the principle of verifiability. Instead, the aim of analysis using the principle of verifiability and the new logic was to establish, as Carnap puts it, “an exact method of philosophy” (1935: 38).³⁶ In doing so, they sought to rationally reconstruct the logical relationships between statements about first-person experience and objective (or intersubjective) statements about the world to better understand the logic of science.

However, as mentioned in §3 above, Stebbing took the logical positivist interest in logic and in language to signal an interest in communication more generally. Or, at the very least, she thought the logical positivist approach to analysis has implications for communication. After all, the principle of verifiability purports to give a characterization of meaning, and, for Stebbing, communication is about understanding a speaker’s intended meaning. Further, Stebbing picks up on the references to communication in Schlick’s “Form and Content” lectures (even if we might now understand those references more along the lines just mentioned). So, I think she would find it odd to emphasize the importance of language in philosophical analysis and to ground philosophical analysis in an account of meaning, while also maintaining none of it is relevant to questions about communication. As such, she takes the implications of the logical positivist picture of analysis for communication—insofar as it collapses the tripartite distinction between understanding, knowing, and analyzing—as fair game for criticism.

This provides the background against which we can see Stebbing set the stage for ordinary language philosophy’s later engagement with logical positivism.³⁷ For Stebbing, communication concerns questions of understanding, not knowledge or analysis. On Stebbing’s view of the logical positivists, given the “slag of historical languages,” the new logic would help clarify what it is we understand when we communicate to one another, revealing that we communicate via structure, not content. Consider, though, the second commitment of Moore’s view of philosophical analysis: “there are expressions in ordinary usage which unambiguously express . . . propositions which we know to be true. A proposition is unambiguously expressed when what is said is understood” (Stebbing 1933a: 7). Such propositions, as expressed in ordinary language, are the starting points of philosophical analysis. The logical positivists, however, “start from a priori assumptions with regard to the nature of language and the principles of symbolism” that they use “to draw limits with regard to what we can think” (36). Stebbing thinks this gets things backwards. She says, “We must proceed step by step, beginning with propositions which we know to be true, not ruling out initially what does not fit in” (36).

These remarks should remind us of Strawson’s (1963) later criticisms of Carnap’s project of explication.³⁸ Very roughly, explication involves the clarification of the unclear concepts of ordinary language by translating it into a formal language in which the logical relationships between concepts are clear. Strawson thinks such a project is misguided. Instead, philosophical analysis needs to start with a close examination of ordinary language rather than assumptions about its unfitness for certain purposes. This is because

typical philosophical problems about the concepts used in non-scientific discourse cannot be solved by laying down the rules of use of exact and fruitful concepts in science. To do this . . . is not to solve the typical philosophical problem, but to change the subject. (Strawson 1963: 506)

For Strawson, we solve typical philosophical problems by examining language as it is actually used in particular contexts and for particular purposes to clarify what is involved when we ordinarily employ sentences and terms successfully.

For both Strawson and Stebbing, then, the aim of philosophical analysis is not to find a replacement concept for our unclear, ordinary ways of speaking that fits neatly into a well-ordered logical system freed from the “slag of historical languages.” Instead of adopting proposals about meaning and verification that rule out ordinary language as unclear or unsuitable for certain purposes, one must start with ordinary language. For Stebbing, this means starting with propositions that unambiguously express what we know to be true.³⁹ This might seem an odd view of philosophical analysis. If we understand some expression, what need is there for analysis? This question was partially addressed towards the end of §2, where I drew attention to Stebbing’s distinction between a statement being unambiguously expressed and it being clear. However, to borrow from Strawson, the “need for the elucidation of concepts . . . can coexist with perfect mastery of their practical employment,” which is something he claims the logical positivist fails to recognize (1963: 509). Strawson’s point echoes the distinction between understanding—“perfect mastery of their practical employment”—and analysis—the “need for the elucidation of concepts”—already present in Stebbing’s earlier work.

On this view, philosophical analysis is aimed at making clear what is involved when we know some statement of ordinary language to be true. For Stebbing, it is important to note that “This investigation is not linguistic” (1933a: 36). Recall her intention-based account of communication: A speaker can intend to refer to the world, to their experience, or even to their own use of words, and their choice of expression makes clear their intended reference. When a speaker intends to refer to the world or to their own experience rather than to their use of words, an analysis of what they said does not only tell us about language; it also tells us something extra-linguistic about the world or experience. A similar point shows up in Austin’s later remarks about the method of ordinary language philosophy.⁴⁰ Like Stebbing, he cautions against thinking that because ordinary language philosophers analyze language, they are merely interested in words. Austin says that

When we examine what we should say when, what words we should use in what situations, we are looking again not merely at words (or ‘meanings’, whatever they may be) but also at the realities we use the words to talk about: we are using a sharpened awareness of words to sharpen our perception of, though not as the final arbiter of, the phenomena. (1956–1957: 8)⁴¹

While Stebbing would not quite put the point as Austin does, she makes clear that even though philosophy begins with truths unambiguously expressed in ordinary language, philosophers are not simply interested in language. Of the logical positivists, she says,

Not only is their conception of analysis defective, but, further, their conception of the kind of facts to be analysed is inadequate. They treat all facts as linguistic facts. Hence, they suppose that the first problem of philosophy is to determine the principles of symbolism, and from these principles to draw limits with regard to what we can think. (1933a: 33)

But Stebbing says, “We cannot clarify our thoughts by thinking about thinking, nor by thinking about logic. We have to think about what we were thinking about” (36). An account of philosophical methodology like that offered by the logical positivists fails by exclusively focusing on linguistic facts. Language is used to communicate our thoughts about the world, but philosophical analysis is not about making clearer the meaning of the sentences by specifying verification conditions. On its own, translation of an expression into a clearer language does not reveal anything about what we were thinking about; it simply replaces one expression with another.⁴² Instead, analysis renders our thoughts about the world, including our experience of the world, clearer by showing us what is involved and must be the case when we know and understand sentences expressed unambiguously in ordinary language. For Stebbing and later ordinary language philosophers, then, it seems the tripartite distinction between knowing, understanding, and analyzing the propositions expressed in language is one way of preserving philosophy’s concern with the world rather than mere words.

6. Concluding Remarks

Unlike Strawson and Austin, Stebbing shared the logical positivists’ interests in philosophy of science.⁴³ Yet, as a matter of historical fact, the logical positivists’ concern to use formal languages to develop a scientific philosophy barely appear in discussions of their work in early and later British analytic philosophy (Richardson 2017). Instead, in large part due to Stebbing’s influence, the logical positivists were taken to be using formal languages for the purposes of philosophical analysis to clarify ordinary language. Moreover, in advancing the principle of verifiability as a theory of meaning, Stebbing took the logical positivist conception of analysis to have untenable implications for communication, particularly in light of their adoption of methodological solipsism. For Stebbing, these implications can be traced to the ways the logical positivists collapse Moore’s tripartite distinction between understanding, knowing, and analysis. And, according to Stebbing, the problems with collapsing this distinction are illustrated by logical positivism’s treatment of statements about other minds and the past.

In reconstructing these criticisms, I have made the case for Stebbing’s central role in the early reception of logical positivism in British philosophy. Further, her criticisms, as well as some of the assumptions about language, communication, and analysis underpinning them, are echoed and developed by later ordinary language philosophers like Strawson and Austin. Stebbing’s early criticisms place logical positivist concerns squarely alongside those of British, analytic philosophers in a way that helps explain why, for example, Strawson and Austin would understand the logical positivists as engaged in similar enough projects of analysis to be worth criticizing.