Relevance as the Moving Ground of Semiosis

Abstract

All levels of semiosis, from the materiality of signs to their contents and the contexts of their application, are structured by a selectivity in human experience and action that foregrounds only a fraction of the situation here and now. Before Sperber and Wilson, concepts of “relevance” were proposed in both semiotics and phenomenology to analyze this selectivity. Building critically on Alfred Schutz’s phenomenology, I suggest that a productive way to capture the fundamental role of relevance in processes of meaning-making is to see relevance as the outcome of an interplay between two antagonistic tendencies. On the one hand, socially stabilized and individually sedimented “types” guide our experience and action along established patterns. On the other hand, we are actively open to new and unexpected aspects; we are ready to deviate from types and to change typical patterns. Only both tendencies taken together account for our semiotic behavior in context. Spatial metaphors such as “ground” illuminate only a part of this interplay. Due to the double movement in what becomes relevant to us, the typical ground on which we produce and interpret signs is constantly being shifted and re-grounded, which keeps driving on an endless process of semiosis.

1. Introduction

“Relevance” is a notion that, long before the well-known relevance theory of Sperber and Wilson, connected research in phenomenology and semiotics, spanning all levels of sign use from the material to the cognitive [1,2]. As can be expected from this range, it is a very abstract notion, but even at this abstract level, it comes with surprising complexities that will require further research to disentangle. Building critically on some general principles of Alfred Schutz’s phenomenological theory of relevance, this paper focuses on one fundamental complexity by suggesting that relevance follows a twofold rather than a single dynamic that drives processes of meaning-making. While I mainly discuss linguistic communication, the prototypical example of sign use, the argument will suggest that natural language cannot be understood without reference to this general dynamic of relevance and, therefore, cannot be separated from processes of semiosis beyond language.

In Section 2, I will use Sperber’s and Wilson’s relevance theory as a starting point and contrast it with the popular view that language has a special place in our semiotic lives because it stabilizes meaning through fixed, and therefore reliable, rules. Wilson and Sperber remind us that concrete linguistic meaning always depends on context, and they help formulate the problem of relevance as that of explaining how we select, mostly without any awareness of doing so, relevant aspects from context, where “context” is understood as a potentially infinite situation of utterance. Nevertheless, their conception of relevance falls short of providing this explanation, as it is restricted by an analytical framework modeled on logic and natural science.

This will make it necessary (in Section 3) to broaden the theoretical horizon. I will point to relevance research done in phenomenology and semiotics since the 1920s. This research confirms that conscious attention to relevance, and with it, everyday notions of relevance, are the tip of an iceberg. Relevance includes a largely automatic selective activity that ranges from utterance interpretation down to the perceptual level of hearing a sound as a word. The systematic character of the latter level raises the question of whether relevance in general provides a reliable ground for communication.

In Section 4, I will look at “ground” and other spatial metaphors that have been used to describe relevance; in fact, the word “relevance” may itself originally be one of them. But while the intuition that relevance is grounded remains important, spatial imagery obscures the irreducibly temporal dimension of relevance.

The temporal dimension is approached in Section 5, where I focus on Schutz’s conception of “meaning” in a wide sense (far beyond linguistics) as selection from an infinite “fullness”, from which Schutz develops his later phenomenology of relevance. His conception highlights an affinity between the conceptual form, including theoretical language, and the experience of space, both metaphorical and literal. Relevance involves a complex dynamic that includes this “spatializing” movement and eludes it at the same time.

This dynamic will be approached more closely, in Section 6, through Schutz’s concept of “typification”, which builds on Husserl’s genetic phenomenology. Types are patterns of meaning which, unlike strictly fixed rules or quasi-spatial structures, remain open to changes. While the repetitive character of types helps ground relevance, their openness allows for flexibility in our encounters with ever-new contexts.

Nevertheless, this openness needs a more thorough analysis, which is attempted in Section 7 and Section 8. While there are good reasons to assume a general tendency in human beings to follow typical patterns, the only kind of openness compatible with this tendency, I will argue, is passive and takes the main form of imposed openness, where the disturbance or interruption of a typical pattern forces us to look beyond our system of types. I will argue that this is not the only kind of openness we display. Our largely unproblematic and spontaneous flexibility and creativity even in everyday life—exemplified in our language use—suggests a second tendency that is often overlooked, a tendency to deviate from typical patterns, an active openness that is opposed to the impulse to follow types.

Finally, in Section 9 and Section 10, I will suggest that relevance is the outcome of an interplay between both tendencies, neither of which can be reduced to the other, as they point in opposite directions. On the one hand, the selection of what is relevant tends to follow typical patterns. This first tendency allows for some openness of meaning, but only in its passive, and, more precisely, imposed form. On the other hand, the selection of what is relevant tends to deviate from typical patterns. This second tendency constitutes an active openness of meaning. In other words, relevance is grounded by types, but this ground is kept in constant motion by the opposed tendency to move away from types. Which aspects of the potentially infinite situation here and now become relevant to us depends on a complex dynamic that involves both tendencies.

2. Against a Skeleton Model of Language

Although semiotics studies far more than language, language has often been attractive as a baseline with which to compare or contrast other kinds of signs—“languages of art” [3], for instance. Words have been the prototypical signs at least since Aristotle (De interpretatione 16a) analyzed them as sēmeia, inviting all of the theoretical difficulties that follow. Furthermore, the special role of language is not just an accident of the history of thought. Our mother tongue is a chief medium through which we take in the world and understand each other and ourselves in a way that is effortless, and yet very precise. Especially in fields outside semiotics proper, such as social philosophy, this role of language is often explained by models that posit “rules” not in the more flexible sense of “rule-following” found, e.g., in the later Wittgenstein, but in the sense of rigid procedures to be identically applied across different contexts [4,5]. A model invoking fixed rules in this sense portrays language as something similar to a skeleton that gives stability and articulation to the softer layers of signs around it, to gestures, faces, pictures, smells, etc. But such a skeleton model has not always been self-evident. Wilhelm von Humboldt, for example, insisted, on the contrary, that we should see a language not as a fixed structure, but as a supple potential that adapts to the situation of use and depends on the productive activity of its users [6],1 which would suggest a precarious dialectic between linguistic structures and the processes in which such structures are produced, reproduced, and changed [7,8]. Only in the second half of the 20th century did it become a widespread assumption that a natural language furnishes us with stable structures or fixed rules and that these guarantee a hard core of meaning that is identical across situations and users. Among the sources for this assumption were the view of language as a system in Saussure’s Cours, the Shannon/Weaver mathematical code model of communication, and the reception of both in structuralist theories. Another source was the impact of Chomsky, who advocated formal structures and methods close to the mathematical and natural sciences.

Over the last three decades, the skeleton model of language has been attacked from within the paradigm characterized by this assumption under the umbrella of “relevance theory”. Cognitive scientists Sperber and Wilson [9,10] do assume that a language can be described as a stable “code”. But even so, they argue, straightforward “decoding” would never be enough to explain how we understand even the propositional core meaning of our linguistic utterances in context. This is obvious with indexicals, such as “you” or “here”. However, Sperber and Wilson adduce a broad range of well-known phenomena, such as ambiguity, vagueness and intended degree of precision, wider or narrower meaning, figurative uses, irony, and “ad hoc concepts” [11], to show that the meaning conveyed by utterances may always be different in new ways from what the “code” on its own would have us expect. Furthermore, given this possibility, even where words are used to communicate exactly what the dictionary says they mean, successful communication depends on the decision not to deviate from standard meaning here and now—a choice the rules of language do not make for us. This dissolves the impression of a firm skeleton of meaning even within language. Any presumptive code built out of fixed rules—always in the sense I just sketched—would both underdetermine and overdetermine the meaning of an utterance in context [12]. Meaning in linguistic communication always depends on context.

However, Sperber and Wilson point out that context is potentially infinite in the sense of the unlimited possibility of invoking—in theory—ever further aspects of the situation surrounding any given utterance. Potentially, a context includes the perceptual environment here and now, what was said before, expectations of what might happen next, general knowledge, knowledge of participants about each other, etc.2 This is where relevance enters the picture. “The” context, i.e., the relevant context, is not “given”, but “chosen” [9] (pp. 132ff.). When we use words to mean something and when we understand what others mean by their words, we intuitively select that fraction of the potentially infinite context that is relevant to the utterance here and now, and we draw on that fraction in the relevant way. For obvious reasons of human finiteness, we do not do this by comparing an infinite number of possibilities and then deciding on one. Yet neither do we follow some rigid meta-code or fixed selection rules, as the intuitively relevant contexts vary from situation to situation in kind and content. The fact that we manage to understand each other’s words in context at least some of the time cannot be explained by pointing to fixed rules, whether these are claimed to exist empirically or introduced by way of theoretical “idealization”, as, e.g., Habermas does [12]. Relevance, then, becomes the title for an enigmatic activity that enables human beings to pick out just the right—i.e., relevant—aspects out of an infinitely rich situation.

Wilson and Sperber remind us that, however central its role in our lives, language has more in common with softer areas of meaning-making, such as gestures or pictures, than is often thought. Furthermore, they show that the ease and precision with which we use our language is not down to a simple process of decoding according to fixed rules, but to flexible and often creative processes of selection geared towards relevance and to the ability to intuit what is relevant to others in a situation. At the same time, the specific concept of relevance that Wilson and Sperber proposed has, in some ways, become an obstacle to further progress in relevance theory.

The immanent problem of Sperber’s and Wilson’s cognitive-scientific framework is that it is too narrow to account for the role of relevance, even in linguistic communication. The two theorists define relevance in terms of cognitive efficiency: A piece of information is maximally relevant if (1) a maximum of “cognitive effects” can be derived with (2) a minimum of “effort”. However, (ad 1) their definition of “cognitive effects” in terms of a logic of propositions and inferences misses entire classes of “effects”, such as goal-related [14], “perceptive” [15], or “affective” [16] ones. Certain aspects of a situation can become relevant to me (and will, therefore, guide my interpretation of an utterance, something that the producer of that utterance must take into account) primarily because they help me pursue and modify goals I have or help me form new goals, because they change my aesthetic sensibilities, or due to their emotional impact. But such relevance cannot be captured in terms of logical inferences that I draw from assumptions about the world, or even from assumptions about my own goals, sensibilities, or emotional makeup. The effects in question operate beyond or beneath propositional logic. (Ad 2) The effects in question may be indirectly included via the concept of “effort”: Maybe it is easier to derive cognitive effects when a practical or emotional connection is present. However, the notion of “effort” that Sperber and Wilson use is defined in terms of natural-scientific ideas about brain physiology. To my mind, no such notion can, in principle, do the explanatory work it is assigned because it fails to answer a crucial question for a relevance theory of communication. Producers and interpreters of signs can communicate successfully where at least one side is able to recognize and foresee what is, or will be, relevant to the other in the situation. Depending on the communicative scenario and the abilities of the communicators, this task can be symmetrical, or it can fall more to one side than the other, and characteristic errors occur where participants lack the requisite abilities, e.g., due to young age or autism [17,18]. However, as long as relevance is defined—via “effort”—as something happening in a person’s brain, the theory cannot explain how we foresee or even recognize what is relevant to others (or even to ourselves, for that matter) [14].3

In sum, Sperber and Wilson show that the problem of relevance crops up even at the heart of a paradigm focused on propositional logic and natural science; nevertheless, their specific concept of relevance indicates that the problem cannot be solved or even fully addressed within that paradigm.

Traditions that are less well-known today might help develop the concept of relevance further. Relevance theories were developed in semiotics and phenomenology (among other disciplines [20]). However, most research done in the Sperber/Wilson framework has kept aloof from both. As for semiotics, Sperber and Wilson needlessly reduce “modern semiotics” to what they call “the code model”, the idea that meaning is communicated by coding and decoding on the basis of a stable and decontextualized system of signal–message pairs [9] (pp. 2ff.), i.e., on the basis of fixed rules. This idea, modeled on parts of Lévi-Strauss’s work referring to the Shannon/Weaver model of communication, hardly resembles any modern semiotic approach, and certainly neither Peircean semiotics nor 1930s structuralism [2]. In fact, the term “code” was explicated in a way compatible with Sperber’s and Wilson’s argument—and based on extensive criticism of Lévi-Strauss—already in Eco’s 1968 Absent Structure [21].4

Phenomenology was cut off even more effectively from a relevance theory that works within a framework restricted to logic and natural science. Phenomenologists have argued that logic and science, like all human enterprises, have their foundation in the relevancies of the “lifeworld” of our action and experience as social, cultural, embodied beings [23,24]. Yet logic and science are productive enterprises precisely because they abstract away from most of this foundation. Propositional logic and natural science cannot provide the framework for a general theory of relevance because, by their very nature, they simply take for granted most of what relevance is. Not only the practical, sensual, or affective dimensions of relevance, but also the mechanisms that connect the individual plane of relevance with its social plane remain invisible unless we step beyond this framework. This is not a step into an unknown void if we make use of earlier relevance research in semiotics and phenomenology.

3. The Rest of the Iceberg

Phenomenologists and semioticians affirmed a massive context dependence of linguistic meaning before Sperber and Wilson brought it back to the table. Husserl’s interest in the foundations of logic and mathematics made him recognize by 1913 that the meaning of practically all our linguistic utterances in everyday life is “occasional” in that it can only be understood in conjunction with the extra-linguistic (e.g., perceptual or social) situation [25]. Paradoxically, Husserl’s ideal of truly context-independent meaning highlighted the fact that the meaning of “all empirical predications” is “occasional” [26] (p. 7), which leaves up to context most of what utterances actually mean.5 The two strands came together to some extent in his later analysis, according to which the merely verbal meaning (Bedeutung) of an expression is only one element in its full and actual meaning (Sinn) [28]. Among other semioticians, Bühler [29], who was influenced in part by Husserl [30], pointed out that real-life linguistic meaning behaves nothing like any ideal of “meaning-constancy” postulated by pure logic, or by what Sperber and Wilson would later call a “code model” of communication. Language, he argued, flexibly adapts to diverse and novel situations, but as a result, the competent use of language requires understanding the situational context. This point was made again by Eco [21] in his early analysis of codes: When decoding a message, the receiver has to choose from a potentially infinite and constantly changing multiplicity of codes and sub-codes, and how this choice works can ultimately be understood only with reference to “circumstances” (la circostanza) of the message that are not (or not yet) coded.

Moreover, the contextuality of meaning was connected to notions of relevance early on. The term “relevance” marks a core area where work in semiotics and phenomenology intersected from the 1930s onwards [1] (pp. 26ff.). The breadth of its application suggests that the problem of relevance is not confined to the meaning of linguistic utterances or even of signs more generally. It also concerns the very perception of signs as signs. In fact, this is one historical starting point for the notion. Trubetzkoy [31] pointed out that, acoustically (phonetically), a spoken word is a complex sound event with a potential infinity of properties. But only a handful of features—depending on the analysis, as little as a dozen—that the sound system (phonology) of some language or other picks out allow us to hear this sound event as a word. All the other properties of the acoustic material do not serve this function and, in this sense, amount to mere noise. Thus, hearing a sound as a word is a feat that involves selecting, from an infinitely rich acoustic event, a few aspects as “relevant”—a term that Trubetzkoy would come to use after an exchange with Bühler [32,33]. Bühler himself [29] went on to generalize Trubetzkoy’s idea to include not only the material basis of meaning-making, but also the dimension of meaning itself.

Such a breadth of application should not be seen as an argument against a unified concept of relevance. Abstraction is necessary if relevance is to help explain how we process linguistic meanings in connection with a wider context that includes, among other things, our perceptual situation. As Sperber and Wilson insist, communication and cognition, including, within communication, a spectrum ranging from saying to showing, must be analyzed along the same lines. It seems only appropriate, then, that Eco [21] uses a notion of relevance of similar generality to refer to the fact that codes selectively determine “relevant” units and features, from perceptual codes to the basic syntactical articulation of messages and their interpretation, in terms of “cultural units” and the connotational values attributed to them.

Furthermore, the example of language shows that relevance is, for the most part, an inconspicuous but powerful factor in our action and experience, which may explain why the problem of relevance is often overlooked. When we explicitly talk about relevance, we are often concerned with the relation of an utterance or a text to something else: Is it pertinent to the topic? Did it answer the question? Does it help us reach a purpose or gain new insight? This is presumably why the word “relevant” first arose as a technical expression in jurisprudence [34] and then became a key term in library studies and information science [35]. However, evaluating the relation of an utterance to something else presupposes that we can first identify and understand that utterance, and as we have seen, principles of relevance are already tacitly at work at these levels, too. Our conscious attention to relevance illuminates the mere tip of an iceberg. This would be consistent with phenomenological theory. Husserl distinguished between the attitude of the mind “living in its acts”, which governs most of our lives, and the attitude of the mind “reflecting” upon its acts (which includes the phenomenologist’s own reflection). Schutz extended this distinction to his concept of relevance: We are always “living in” relevancies, even when we are not “looking to” them [36] (pp. 66f.), cf. [37] (p. 182). This is clearly true of language once our concept of relevance is sufficiently wide to include the linguistic patterns that users follow without even knowing them, an example already used by Kant to argue that our perception and action are governed by tacit regularities [38] (p. 76).

At the same time, as Eco [21] notes and as would increasingly worry him in later work, rather different dynamics seem to be at work when we move from one level to another. The repertoire of acoustic features that a language’s phonology picks out as relevant remains the same for long times, sometimes over centuries. At the other end of the scale, the meaning of a word in a sentence, or the meaning of the whole sentence, can change in a matter of seconds as different contexts become relevant to its interpretation. While the relevancies that shape the material basis of language plausibly follow something similar to stable rules, the relevancies that underlie our interpretation and production of meaning through language depend more strongly on the situation here and now.

This is a problem if we aim to understand how communication works. What makes the notion of fixed rules so popular is that they provide a simple way to do so. Rules that remain strictly identical across time would make for easy and reliable orientation, and if they are shared among individuals in a group, their reliability can be shared as a safe route (barring disturbances) for group members to understand each other in any given context. In essence, this is the idea behind a view of language as a skeleton, to which authors like Searle and Habermas continue to adhere. This may be an apt description of what happens on the phonological level: Barring disturbances, speakers of the same language have no difficulty hearing wildly different sound events as instances of one and the same word. In contrast, the interpretation of utterances can vary with the situation and even the individuals involved, and if there is no analog to the phonological system on this level, it remains unclear what prevents meaning in context from becoming unpredictable, unreliable, and, perhaps, even random. Given that we are quite good at understanding each other through language in context, the selection of what is relevant here and now cannot be a random matter.

4. Grounds and Maps

If relevance does not float freely in space, it must be grounded in some manner. In his late phenomenology, Husserl says that our “lifeworld” provides a common “ground” (Boden) for us to understand one another. But this metaphor raises more questions than it answers. Adopting Husserl’s parlance, Habermas claims that the structures of a natural language are sufficient to anchor each speaker in the “ground” of a shared lifeworld [5]. He also re-interprets the lifeworld in terms of Searle’s “Background” [39], which is needed to supplement linguistic meaning in context, but is itself “functionally equivalent” to a language understood as a “system of rules” [4] (p. 142). However, the findings about relevance in context cited above contradict these claims. Large portions of the ground we are looking for not only lie outside language, but also transcend the simple rule model that Searle and Habermas have in mind. Peirce’s earlier concept of “ground” seems more promising in this respect. It refers to features and relations, whether in the world or in our experience, that go beyond pure semiotic convention and help tether signs that would otherwise threaten to drift freely through contexts. The relative flexibility of this notion—compared to fixed rules—has allowed it to be put to different uses [40,41].

One thing that connects notions like “ground” and “background” with that of “relevance” is their spatial imagery. A ground is a reliable terrain to stand and walk on. A fixed lay of the land both shapes and restricts movement, and it opens and imposes views of the landscape. “Relevance”, too, is a metaphor taken from the realm of solid-state physics. In its original, juridical application, “relevant” referred to arguments or evidence that “weighed” enough to press one side of the scales of justice towards the ground and, thus, to visibly “lift up” (relevare) the other side. Phenomenologists have proposed a different, but also spatial, image to elucidate relevance: What is relevant “stands out” by itself from the “background”; there is something like a spatial “relief” (a word that ultimately stems from Latin relevare, too) that imposes this figure–ground relationship on anybody who observes it [42] (pp. 56f.) [43] (p. 102). It is tempting to connect the metaphor of “ground” with that of relevance as “relief”, as has indeed been done. Schutz not only uses Husserl’s view of the lifeworld as ground and the gestalt-theoretic language of figure and (back)ground, but he also compares relevance to a “landscape”, a topographical relief with mountains, valleys, and saddles. More precisely, he refers to a map or a three-dimensional model reproducing the “profile” of such a landscape through “contour lines” (“isohypses”) that indicate different heights by connecting points of similar elevation [44] (p. 93) [36] (p. 34).6

One might want to expand the spatial imagery by saying that our conventional signs are “built” upon a pre-existing ground that conditions their design and stabilizes their use. After all, many terms associated with signs, such as “structure” or “system”, are originally metaphors from the field of construction work and architecture. Bühler, for instance, writes that Trubetzkoy “erected a phonological building” (Gebäude) after drawing the indispensable “basis” for this building: phonetic analysis [32] (p. 23). Moreover, he describes this “basis” using the same comparison as Schutz: as a “plastic relief” with “contour lines” that represent a “mountainous landscape” (ibid., p. 32f.).7 Certain sound features are treated as “relevant” for distinguishing words in a number of different languages, and this has to do at least in part with the fact that the physiology of the human vocal tract and hearing apparatus makes these features “stand out” among others. Perhaps, then, we might ask: Do those aspects of the situation surrounding an utterance that are relevant to interpreting this utterance here and now stand out due to the “topography” of our minds or the world itself?

We should be cautious with these metaphors. While Bühler applies the landscape metaphor to the material “basis”, to the level where phonetic analysis shows certain sound features to be statistically more frequent than others, he stresses that different languages can treat quite different features of the very same sound as “relevant” within their phonological system. While Schutz does speak of “structures” and “systems” of relevance and connects these to “relevance lines” in a “relevance landscape”, what he talks about—like Bühler—is the representation of a landscape and the activity of a “cartographer” [44] (p. 93), not the landscape itself.

The relation between a map and its referent raises classic difficulties. In a Kantian vein, we might say: If our relevance “map” is foundational not only to our use of language, but also to our attention and perception, it is hard to see how we could ever experience the actual landscape beyond our map of it, and we might also wonder in what sense such an experience would even be relevant to us. However, if we think of the landscape in terms of real risks and opportunities, this would put us in the precarious situation of a driver who keeps looking at the map rather than the road. Alternatively, if we assume that our map is perfectly accurate, that it reflects the real relevancies of our surroundings, we run into another difficulty. In different situations, ever new aspects of the world can become relevant, however far-fetched and at whatever degree of detail and shading. In Husserl’s language, we may be motivated to enter indefinitely into the inner and outer “horizons” (another geographical metaphor, referring this time to the perspective of an observer placed within a landscape) of any topic. To reflect this possibility, our relevance map would need to replicate the landscape like the map Borges imagined—and would be equally difficult to handle [46] (p. 146).

Even worse, this perfect map would resemble neither a map nor a landscape. The whole idea of a landscape was introduced to illustrate a stable relevance profile, but if what is relevant changes with the situation, an accurate map would have to overlay an infinite number of different possible profiles and end up as a black piece of paper that would convey the same cartographical information if left blank. Eco argued that an “ontological” structuralism, when brought to its logical conclusion, destroys the very concept of a structure: For every concrete description of a structure, a still deeper structure must be assumed that contains this description and possible others as partial and, thus, imperfect realizations of itself. The vanishing point of this logic is not a structure, but a quasi-Heideggerian nothingness that contains all possible structures but cannot be identified with any of them [21]. A perfect relevance map seems to share this fate by destroying the concept of a map. For the idea to have any meaning, therefore, a relevance map cannot refer to all possible situations, but must be confined to a particular situation and a viewpoint within it. This is indeed how Schutz sees it: “Relevance lines (isohypses) are always drawn from the Now and Thus”, and the landscape in question is “the landscape of my Now” [47] (pp. 54, 121).

In essence, then, the idea of a relevance landscape leads us back once again to the relation between an infinitely rich situation and our selection of the relevant aspects within it. This does not invalidate the intuition that relevance is grounded, or that we are on common ground when we understand each other by selecting what is relevant here and now. Nevertheless, trying to translate this intuition into spatial imagery has suggested that we are not dealing with pre-existing, unchanging structures, but with something more flexible. “Ground” should be taken seriously, but not literally. If spatial metaphors have failed us, we should look more closely at the temporal dimension of relevance and its role in meaning-making. Such an analysis might then also help explain why spatial metaphors like these are so widespread (yet perhaps not inevitable) despite their misleading character. A useful starting point for a temporal analysis is Schutz, who stated that “the problem of meaning is a time problem” [48] (p. 93, original emphasis).

5. Meaning as Process

In the 1920s, Schutz started to analyze the concept of “meaning” (Sinn), which was seen in contemporary interpretive sociology as fundamental to a social world, in a long manuscript unpublished during his lifetime and entitled Forms of life and the structure of meaning [49].8 At the time, Schutz had not yet discovered the significance of Husserl’s phenomenology to his project, and his approach was based instead on a critical reading of Henri Bergson’s philosophy of life. According to Schutz, not only concepts, but also perceptions, actions, and even the unity of one’s own felt body are meaning “constructs” (Gebilde might also translate as “formations” or “figurations”). That is, they are phenomena constituted through the selective foregrounding of certain aspects and relations within the continuous stream (durée) of lived experience that Bergson had described. “Meaning”, according to Schutz, is the “tension” that exists between these selective constructs and the inexhaustible “fullness” out of which they are carved. This complex “structure of meaning” is closely related to the problem of relevance as I described it earlier, which becomes Schutz’s central problem: “Evidently [these are] correlative concepts: meaning relations [Sinnzusammenhänge] only exist between that which is relevant; only that which stands in a meaning relation is relevant. However, an unclarified concept of relevance is presupposed here” [50] (p. 49).

The Bergsonian framework suggested that meaning relations and, thus, relevance should be approached primarily as a matter of time rather than space. In fact, it suggested that the experience of space was itself primarily a temporal matter. According to Schutz, the character of spatial extension accorded to material things, to the ground beneath my feet, and even to my own body as a thing in space is itself a meaning construct, a phenomenon in which certain aspects within the flow of experience that remain the same or repeat themselves are selectively stressed while innumerable changes in and around the objects thus circumscribed are neglected.

This would explain why we like to think of concepts and conceptual networks in spatial terms. For Bergson, our everyday experience as a whole “spatializes” the stream of lived experience, not merely through our literal orientation towards space and spatial objects, but also through concepts, which likewise tend to foreground certain aspects of experience that remain the same or repeat themselves across occasions while abstracting away from changes. As a result, any conceptual analysis of meaning-making, including phenomenological or semiotic analysis, has—already qua conceptual analysis—an affinity for spatial models such as Schutz’s own idea of a relevance “landscape”. While such ideas are often helpful, they should not lead us to forget that meaning-making is primarily a process in time. When comparing relevance to a jigsaw puzzle, the late Schutz notes: “[…] all of this has been described here in spatial and therefore inadequate metaphors what are [really] hardenings, products, results of processes of the stream of consciousness in the Bergsonian durée” [51] (p. 203).

“Hardenings” and “stream” are, of course, once again metaphors (somewhat reminiscent of Hjelmslev’s “magma”) taken from the realm of things in space. The most fundamental notion in Schutz that describes processes of spatialization without already using spatial concepts is “repetition”. Given Schutz’s reading of Bergson, it is clear that when we say that “the same” repeats itself, a host of changes in experience have been abstracted away. No experience is, therefore, strictly speaking, the same as a previous one, Schutz stresses, and even those aspects that are repeated undergo the subtle Kierkegaardian changes that separate the reconfirmed or prolonged experience of something from the first time around: “Being recurrent, the recurrent is not the same any more” [52] (p. 115). In this sense, meaning is unlike a thing in space with its (supposedly) unchanging and self-sufficient identity, and presumably, any “grounds”, “structures”, etc. that help articulate meaning are unlike their literal counterparts in the same way. This would fit our experience of the material basis of language even though the phonological system comes close to the ideal of a fixed landscape. We do hear two utterances of the same word by different speakers or by the same speaker at different times as utterances of one and the same word, but we are equally aware that these are different, indeed unique sound events where only certain relevant features are repeated, unlike with an identical thing persisting through time.

Nevertheless, Schutz did not share Bergson’s dualistic assessment of the matter. For Bergson, our everyday “spatialization” distorts or impoverishes the original reality of “life” that is only accessible to meditative intuition. For Schutz, in contrast, while meaning is no more than a selection from the richness of experience, this selectivity produces all the reality we can possibly have. Far from distorting the potential fullness of life, the structure of meaning constitutes “forms of life” (Schutz’s use of the term Lebensformen predates Wittgenstein’s) that enable us to experience lived reality in the first place. For Schutz, a pure Bergsonian durée, an unbroken stream of life that contains within it the totality of all possible experiences, is itself no more than a theoretical concept because human experience is by necessity selective. This complex relationship between meaning and durée is what Schutz calls a “tension”. Meaning constructs are mere selections, and they gain all their material from a fullness that contains potentially infinite alternatives—but they are the only form in which this fullness can enter our experience.

Given this tension, the processes through which meaning constructs are formed, reproduced, related to each other, and modified take center stage. Any meaning construct leaves out alternative possibilities, and among these might be aspects that will later prove to be important, perhaps even vital, in some way or another. Since later meaning constructs offer the opportunity to focus on aspects that were missed before, the dynamic of the processes in which meaning constructs motivate and follow one another in time becomes a core issue. One way to approach the problem of relevance is to focus on this dynamic.9 To describe it, Schutz used another concept that he initially adopted from sociology and later found confirmed and elaborated in Husserl’s work: “types” or “typification”.10

6. Types

With respect to the problem just raised, Schutz’s concept of a “type” (Typus) [24,48] performs a double duty. On the one hand (1), it captures a tendency within experience to establish patterns of meaning and to continue following these patterns. On the other hand (2), a merely typical pattern is flexible: Unlike a strict rule or a fixed structure, it allows for exceptions and atypical cases and remains open to changes in the pattern itself.

Ad (1): Thanks to the first of these two characteristics, a type can roughly perform the function of “grounding” relevance in the sense that I sketched earlier. By following established paths of repetition, my meaning constructs immediately pick out for me, from ever new and complex situations, certain aspects that have so far proven viable, helpful, or at least not harmful. Thus, types help constitute a world of familiar objects and events that I can recognize and put to typical ends using typical procedures. Where the relations between these typical elements are themselves typified, they form complex “systems” in which one element routinely calls forth certain others. Most importantly, since types can abstract away not only from variation across different contexts, but also from differences between individuals with their idiosyncrasies and personal situations, they enable us, through language and other signs, to share a cultural common ground and learn from the experiences of others. It is clear that, like social rules, many of these types will have normative, sometimes even coercive, force.

Ad (2): However, a no less important characteristic of types is what distinguishes them from a literal ground or even from strict rules. Against the background of Schutz’s conception of meaning as a tension between highly selective constructs and a fullness of experience that contains potentially infinite alternative possibilities, it is clear that a maximally typical meaning may not always be what is most appropriate in a situation. In the course of experience, “novel” situations may always arise for the first time that do not fit the system of types presently at our disposal. The atypical situation may force us to adapt our types or to try and establish new ones. It may make us realize an aspect that our system of types had until now neglected but which now calls into question the validity of this system. Furthermore, even where we encounter an unproblematic and typical situation, we know that it is not strictly a repetition of the same, but a unique situation that may require modifications or additions with respect to our typical patterns. As a result, even where types have been confirmed by generations and are “taken for granted ”, they may, in principle, be called into question at any time; they are valid only provisionally, only “until further notice”, as Schutz [24] (p. 7) puts it.

The concept of a type goes some way towards explaining how we select, from a potentially infinite context, what is relevant here and now. Typical patterns of meaning (always in Schutz’s sense) articulate, in a reliable and systematic way, our way of coming to terms with a concrete situation. At the same time, the flexibility and openness of types allow us to react to novel situations and to take into account that every situation is unique. A shared repertoire of types and a shared sensitivity to what is beyond the purely typical helps different individuals, especially members of the same group, pick out what is relevant with a far better than random chance that their choices will coincide. This, in turn, enables them to pick out the aspects of context that are relevant to interpreting a given utterance with a good chance of agreement and, thus, of successful communication. There is no guarantee of such agreement, and misunderstandings will occur. However, they will not be ubiquitous, as they would if relevance were a random matter. In principle, anything at all might become relevant in a situation, but even where atypical aspects become relevant to people in concrete situations, these aspects are usually related to existing types as modifications, exceptions, wider or narrower versions, negations, etc.

The double duty of types is quite visible in language. On the semantic and syntactical level, natural language is the “typifying medium par excellence”, as Schutz puts it, a “treasure house of ready made pre-constituted types” [24] (p. 14). Nevertheless, meaning in context is not a robotic application of these types. Instead, it is related to them through the various transformations of presumable dictionary-type meanings that were mentioned above in connection with Wilson’s and Sperber’s relevance theory. The same seems true of the material basis of spoken language. As Bühler notes, while the phonological system makes only certain aspects of sound events relevant to their recognition as words and abstracts away from all others, it is open to many other aspects that may become relevant depending on the situation [32] (p. 48f.), such as a male or female voice, a regional or individual accent, or an emotional tone. These relevancies, we might add, are not random, but usually related to types as instances, modifications, etc.11 In sum, as Sonesson has argued [58], the concept of types has systematic significance to semiotics even beyond, for instance, Eco’s “cognitive types” [38].

Furthermore, the concept of types has epistemological consequences when applied to the standpoint of the scientific observer. What helped Schutz develop his two-sided notion of types was the late Husserl’s concept of “empirical certainty” or “presumptive certainty” as a certainty that remains permanently “on notice” (auf Kündigung) [59] (§ 77). Yet for Schutz, the concept of types plays a more central role that even includes the work of the researcher him or herself. For Schutz, phenomenology deals mostly with types rather than the kind of invariable structures that Husserl called “eidetic” [60]. The objective reality we live in is constituted through intersubjective—and semiotically mediated—processes in which typical patterns of meaning are established, unified, handed down, and negotiated within existing cultural and historical frameworks. Likewise, the subjective perspective of every person is deeply shaped by such processes going on in the social world into which they were born and in which they live. This means that the material investigated by phenomenologists will usually have the form of culturally and historically specific types rather than timeless and universal essences.

Moreover, phenomenology, like all science and scientific objectivity, is itself based on the pre-scientific lifeworld. While the phenomenologist abstracts away from many aspects of everyday experience and action in search of general traits or properties, this is formally still the same operation that people perform in everyday life when acquiring and testing types. Even when phenomenologists are certain to have found a truly universal feature of consciousness, their certainty is “presumptive” in that they should be ready to revise their claim if further investigation reveals cases that call it into question. Hence, even when phenomenologists describe “objective” structures underlying relevance, they do so by abstracting away from their own relevancies as individual human researchers guided by the relevancies accepted within their field, and they do so at the risk of having to revise their claims later.12

A relevance theory that reflects this fact might ultimately lead to the conclusion that Eco draws in the form of a “methodological” structuralism which avoids the pitfalls of an “ontological” one: The presumption that there is something real behind the models posited by the semiotician is entirely legitimate, but semiotic models, like other scientific models (including the models of social science that Schutz [18] wrote about), “must have the courage to consider themselves provisional [provvisori], even when they have allowed operations that were crowned with success” [21] (p. 372).

7. The Passive Openness of Types

The openness of types, with their admittedly “provisional” character, including the types and typical procedures that shape scientific knowledge, is consistent both with Peirce’s idea of semiosis as an endless process and with Husserl’s idea of phenomenology as an enterprise that must be carried on by generation after generation of researchers without ever reaching a definitive end. Whenever I bring a system of types to a new situation, I should be ready for the objection, articulated by the matter at hand or by other people who are involved with it, that something else is relevant here that is not contained in my types and that might force me to adapt them or even to revise them permanently.

Nevertheless, we need to take a closer look at how this openness works. According to Schutz, human beings tend to go on relying on whatever stock of typical knowledge, expectations, and routines they have acquired, and they are justified in doing so because their system of types has stood the test of time. Most of the types we use were tried by generations before us and transmitted as part of a cultural heritage. But, Schutz emphasizes, even types we “take for granted” as “self-evident” are valid only “until further notice”. Even types that go “unquestioned” may be questioned at any time [24]. Yet when and why does this happen, given that a movement away from established types goes directly against our tendency to continue along typical patterns? Schutz did not engage systematically with this question before his early death in 1959, but he mentions several possibilities across his texts. We can choose to go beyond our present types or even to doubt them for independent reasons. Crucially, other people may show us, by example or comment, that the straightforward application of certain types to a given situation is a mistake. Not least do we encounter “novel” phenomena that disappoint our typical expectations or that simply do not fit with any types at our disposition [2]. These are all real and important motivations to deviate from typical patterns, and there may be many more. Still, we might wonder about the underlying dynamic that makes them possible.

We tend to follow typical patterns, and the more we follow them, the more they become ingrained through the processes that Husserl calls “sedimentation” (a metaphor from geology, referring this time to the formation of “ground”). Moreover, through progressive readjustment in use, types should, over time, form coherent systems within each user and become uniform across users within a community. However, if this “conservative” tendency (the word being used here in a purely formal rather than political sense) were the sole dynamic that shapes our experience and action, it would be hard to understand what would motivate deviations from the typical course and why cases such as those mentioned by Schutz arise at all. If my system of types incorporates typical ranges of useful or justified interpretation and typical occasions for doubt, it remains unclear why we choose to go beyond the system, let alone doubt it. If systems of types are uniform across members of the same community, it is unclear how—outside of relations that are themselves typified, e.g., in education, party politics, or the division of labor—other people come to disagree with my types. Moreover, even if they did (for instance, because they are from another community), I might not understand their view or even fail to notice that it is different from mine because my system of types guides my interpretation of other people’s behavior, including my perception of their attempts to communicate with me. This scenario of a ubiquitous “tunnel vision” [64] is, of course, a reductio ad absurdum; it clearly contradicts reality as we know it. To be sure, tunnel vision does occur in many areas of everyday life, which testifies to the power of the conservative tendency behind types. Sometimes, interlocutors persistently talk past each other without even noticing their misunderstanding. On the phonological level, a native German speaker may find it impossible to pronounce and even to hear the purely quantitative difference between Japanese vowels [a] and [a:]; conversely, the qualitative difference between the German [y:] and [ø:] may be inscrutable to a Japanese speaker. On a grander social and political scale, entire societies may for decades overlook ecological consequences that lie at the margin of their typical interests, discourse, and economic goals [65]. Nonetheless, tunnel vision does not have us in its grip everywhere, and even where it does, it can be overcome. The question is how this openness of types is possible.

If we continue to assume, for the sake of argument, that the tendency to stick with established types is the sole underlying dynamic, to my mind, only two kinds of openness are compatible with it, both of which might be called “passive”.

(1) The first is of the classical tabula rasa kind: Phenomena would impress themselves upon us as if on a smooth wax table. We would be passive experiencers of the phenomena in that the process of impression would not involve any activity on our side. If this were possible, it would make us open to phenomena even when they elude our typical patterns of meaning by falling outside the boundaries of our types, through the gaps in them, or below their thresholds (note that I am once again using spatial metaphors here that should not be taken literally).13 However, it is not clear that such openness is possible. According to Schutz, meaning is a tension between selective constructs and an infinite “fullness”, and since all experience is formed through meaning, that fullness cannot be experienced as such. After all, if there is no experiencing activity on our side, in what sense would this still be an “experience”? If there is an infinity of possible ways to make sense of a situation, who or what selects and constitutes the specific phenomenon that somehow happens to us? Even if we do not dismiss it out of hand as a pre-Kantian anachronism, a phenomenologically and semiotically unmediated contact with reality is difficult to conceive. Tabula rasa openness seems to belong to speculative metaphysics rather than either phenomenology or semiotics.

(2) The second kind of openness compatible with a purely conservative dynamic might be called imposed openness.14 When our types face resistance or difficulties, we are forced to stop and think, to revise the type in question, and to solve the problem. Even when the complication (which may be an external obstacle or an internal issue, such as a conflict between incompatible types) arises beyond the radar of our typical experience and expectations, it can still make itself felt indirectly by disturbing or interrupting the typical course. The disturbance then triggers a search for its cause, motivating us to break free from tunnel vision by looking actively beyond our types. This crisis scenario is prominent within the tradition of philosophical pragmatism (even though, of course, it has a longer pedigree, e.g., in Hegel). In Peirce, it appears in his analysis of Secondness as “resistance”, “shock”, or “struggle”, and in the key role that the external “breaking” of the habit of belief through “surprises” plays in his logic of abduction. Eco’s early semiotic theory revolves around “contradictions” and refers to a dialectic between a “closed semiotics” of the code and an “open model” of the message that is “waylaid” by “circumstances” of communication that “break in to disturb [irromperà a turbare] the life of the signs” [21] (pp. 415f.). The scenario appears in phenomenology, too. Max Scheler, for whom Peirce was a central reference, sees “resistance” as the core relation through which human beings in their daily lives encounter reality [68]. The early Schutz similarly takes up the idea of material reality as “resistance” from Bergson’s pragmatism, and later, when further developing his concept of relevance, sees a crucial role of “problems” that force us to revise typical relevancies [53]. As Schutz notes, the Greek problēma, like the Latin obiectum, originally means something that is “thrown before” me [36] (p. 26). The more recent approach of Lohmar, who, like Schutz, elaborates and strengthens Husserl’s concept of “types”, also points to the disappointment of typical expectations as a chief source of novel experiences [67]. But is that the only kind of openness we display?

8. Active Openness

Like the phenomenon of tunnel vision to which it is closely related, imposed openness is real and important in everyday life, and it is frequently what breaks the spell of our conservative tendency to follow types. However, this is not the only kind of openness we exhibit. Just as often, we show an openness in everyday life that lacks any sense of being forced to stop and think, or of being confronted with a problem and having to search for a solution.

In linguistic communication, for instance, recognizing a misunderstanding or interpreting a garden-path sentence like the classic “The horse raced past the barn fell” comes with a discernible feeling of puzzlement and effort (which is reflected in measurably longer reading times, or in back-and-forth eye movements while reading such a sentence). However, this feeling is absent from large stretches of our speaking and listening (or reading), even where we deal with “atypical” language use, such as figurative, loose, or indirect meaning, irony, or ellipsis. If the typical (lexical or encyclopedic) meaning of words were simply “taken for granted until further notice”, that is, until we face some difficulty or counterevidence, the processing of such utterances would follow a route similar to the garden-path sentence: We would start with a standard (literal, truthful, strict, simple, unmarked, etc.) interpretation, which would lead to the dead end of an odd overall reading, which, in turn, would force us to backtrack and revise our interpretation. This scenario has indeed been proposed by some philosophers of language to explain how we deal with non-standard meaning. In contrast, Wilson and Sperber argue that our processing in such cases follows the same path as in standard cases and does so with the same ease: We simply use the first interpretation that is most relevant to us in the situation without having to think about different interpretations or to choose between them [69]. While we need not adopt Sperber’s and Wilson’s specific concept of relevance, phenomenological intuition appears to be on their side. Understanding a mildly creative metaphor or an ironical shift in meaning feels no less smooth and unproblematic than understanding a straightforward and matter-of-fact utterance. Likewise, the production of such atypical meaning in conversation happens, in large part, quickly and unthinkingly, whereas a model restricted to type-following and to the imposed openness that type-following allows for would fall short of even explaining why we produce atypical meaning at all.

Furthermore, the conundrum is not restricted to cases that seem to defy types by being more or less conspicuously atypical. Between atypical situations and maximally typical ones lies the wide field of cases where types are applied with slight modifications, but still smoothly and often without any awareness that we made any modifications at all. An automaton that follows types unless it runs into a problem would be unable to do this. Finally, even the maximally typical situation is never quite perfectly so, if we remember Schutz’s point that types merely abstract away from the fullness of situations and, therefore, do not, in a strict sense, represent “repetition”. Even where I apply a type in a straightforward fashion, what I encounter may not be a “novel” situation, but it is still a “new” situation that, although “familiar as to its type, […] is strange insofar as it is atypical in its uniqueness and particularity” [36] (p. 56, original emphasis).

To capture the dynamic that operates here, a simple remedy is to assume a second tendency that points in the opposite direction of the tendency to follow types: an inherent tendency of our experience and action to move away from the course pre-delineated by types. This tendency represents an active openness, as opposed to the two passive forms compatible with pure type-following and discussed in the previous section. Unlike imposed openness, active openness is not restricted to cases where the typical course is disturbed or interrupted in some way. Unlike tabula rasa openness, it conforms to the Schutzian structure of meaning: Certain aspects are selected out of the potentially infinite fullness of a situation. Thanks to active openness, the activity of selection is free to stray from typical patterns. This second tendency would account for the “independent reasons”, mentioned earlier, for consciously choosing to go beyond our types or to doubt them. Furthermore, it would account for our flexible but largely unthinking use of types just described if we assume that the tendency to move away from types is a natural human impulse that does not require conscious effort and happens just as automatically as the opposite impulse to follow types. As I noted, our conscious attention to relevance illuminates the tip of an iceberg, and nothing keeps us from assuming that an interplay of these two tendencies is at work both above the waterline and below.

While Schutz does not explicitly assume what I called “active openness”, some elements in his phenomenology point in this direction. For instance, he claims that human beings are fundamentally motivated by “anxiety” (Angst), another Kierkegaardian notion, which, unlike “fear” of typified dangers, is directed beyond typical frames [70]. Schutz also uses Leibniz’s notion of “spontaneity”, which could be formulated in our context as an impulse to select ever new and ever different material in our meaning-making processes [64]. However, a tendency in our experience and action to actively transcend typical patterns is evident in many forms. One is boredom: Following the same routine for a long time becomes boring; we start looking for distractions, our thoughts wander off, doing something else would be fun—and before we know it, staying with the routine no longer feels natural but demands increasing effort, or even external coercion.15 Another form—found already in non-human animals [73] and so irresistible that it is known to have killed the proverbial cat—is curiosity, which attests to the intrinsic relevance of whatever is new to us.

Despite being visibly at work, this tendency has often been overlooked, not least by Schutz himself and even more so by his students.16 One reason may be that the functional differentiation of modern societies has tended to reserve a spirit of creativity, curiosity, and discovery to special areas like art and science while creating the residual category of ”everyday life”, formed in the 19th century with an emphasis on the routine procedures reinforced by industrialized labor and rationalization that seem to make every day a repetition of the same [74]. The discourse of modern art, for instance, with its meta-types that demand the constant rejection of old types, outshines the more modest forms of creativity and curiosity that pervade our “everyday” lives [75].17

Another potential reason for why active openness is often neglected is that it may seem unsystematic or inelegant to assume a tension between two opposing tendencies at the foundation of human experience and action rather than dig deeper for a single root that ultimately grounds both. However, all this spatial imagery of roots and grounds might simply be misleading us. If a temporal view of meaning and the assumption of a complex interplay of two tendencies helps explain how relevance works in meaning-making, we should not try to idealize it away in favor of an allegedly more solid foundation.

9. Relevance as an Interplay of Two Tendencies

The suggestion, then, is to conceive of relevance—at the abstract level of the fundamental dynamic of meaning (in the wide sense of Schutz’s Sinn) discussed in this paper—as an interplay between two tendencies that stand in a relation of tension to each other. On the one hand, we tend to follow typical patterns of meaning, which allows for passive and, more precisely, imposed forms of openness.18 On the other hand, we tend to move away from the typical patterns. Such active openness cannot be reduced to the tendency to follow types, nor the other way around, as the two are defined by opposite directions. Which aspects of the potentially infinite situation here and now become relevant to us depends on a complex dynamic that involves both tendencies.

Relevance thus conceived would not be random or unforeseeable. On the one hand, relevance tends to follow a network of typical patterns, and this does provide us with something like the metaphorical “ground” that is largely reliable and is shared with others within social groups. On the other hand, relevance tends to leave the patterns, constantly shaking and warping the typical ground in a movement that might be called anti-typical, were it not for the fact that this movement keeps the typical ground from becoming a free-wheeling abstraction, reconnects it with the fullness of the situation here and now, and, in this sense, re-grounds the typical ground. Crucially, if this second impulse is a basic phenomenological category (and, more precisely, a category within a phenomenologically based anthropology [72]) on a par with the opposite impulse to move within typical channels, it is shared as well, and can be reliable in its own way. We do not usually behave like “cultural dopes” [76] who blindly follow typical patterns without regard for the concrete situation, and we know that others do not either. We often find the same aspects of a situation boring, funny, uncanny, strikingly new, or curious and will, therefore, immediately know what the other person means when they say: “What was that?”. All the while, as already noted, our escapades in the boundless realm of the atypical are not completely free or erratic but relate to our established types, e.g., as modifications or negations of them. In this sense, “[t]he problematic emerges on the foundation [Boden, Grund] of the unproblematic, and the unknown refers to the familiar”, as Schutz puts it [36] (p. 124), but it should be added that what is “problematic” often “emerges” not because it is imposed upon us from outside, but because we are always ready to problematize things.

Relevance as a moving ground, then, would be a basis for the public use of our typical repertoire in context to select certain aspects of the situation as relevant and, thus, to join others in “defining the situation” (as Schutz liked to put it using a formulation by W. I. Thomas) in which our signs are produced and interpreted. Although this process taps back into the fullness of the situation that contains infinite alternative possibilities of meaning as well as into the “consubstantiality” [77] that connects all signs and sign users with each other, it remains a process of selective meaning-making. We neither have a mystical vision of the world as a whole nor can we ever be sure that our agreed selection is objectively the most relevant, or even that our individual selections coincide. The moving ground of relevance never guarantees successful communication; miscommunication, including persistent “tunnel vision”, may always occur. This would make it rational for us to deviate from established types even if we did not already have a natural tendency to do so. To test the typical ground, we must keep moving and re-shaping it, driving semiosis as an endless chain of ever new interpretants.

The suggestion does not seem alien to semiotics. For instance, I have indicated that Eco’s [21] “dialectic” between “closed” systems of codes and an “open” model of message interpretation in context, according to which codes are chosen and applied in response to the potentially infinite “circumstances” of communication, can be rather naturally interpreted in this way. Relevance as a moving ground accounts, on the one hand, for the reliability and regularity of meaning-making processes and, on the other hand, for their flexibility and creativity. It would also reinforce Eco’s ideal of a “methodological” (rather than “ontological”) structuralism that remains consciously “provisional”. Moreover, the idea of a fundamental tension can already be found in semiotic research. When Goodman asked what makes us choose one code rather than another in a concrete context and what guides its application, a question he had explicitly avoided in his Languages of Art, his answer was that we aim for “rightness” [78,79]. As in Eco’s theory and as in phenomenology, this level is distinct from that of truth and falsity. Goodman’s and Elgin’s “rightness”, like relevance,19 involves a tension between two opposed tendencies they variously call “habit” and “progress”, “inertia” and “initiative”, or “entrenchment” and “novelty”. While I focused on the use of language, this tension can be found in many other areas, including such visceral levels of selection as our appetite for specific foods [80], and we may, in theory, expect it to be at work in all areas of sign use where the concepts of relevance and typification have been proposed as crucial, e.g., in visual semiotics [1,57] and cultural semiotics [58].

The high level of abstraction at which I discussed relevance seems to me a prerequisite for developing a general theory of relevance. If relevance has to do with the potentially infinite richness of situations from which relevant aspects are selected, the concept of relevance cannot be made more concrete, at least not from the outset, by limiting it to a specific area. One might object that this abstraction keeps a general theory of relevance from making precise predictions about what will be relevant to somebody in a given situation. However, this criticism would miss the mark. For one, given that potentially anything can become relevant to, e.g., the interpretation of an utterance, asking for a theory that could somehow predict these infinite contingencies would be asking for the impossible: a complete theory of the circumstances of the message [21], the kind of “study of everything” that the later Chomsky rightly rejected [81]. More importantly, this criticism misses the nature of relevance by neglecting active openness. A theory is built on concepts and conceptual abstraction, while relevance involves flexible and often creative processes. In Bergson’s expression, to develop a theory is to “spatialize” our experience by focusing on those aspects that remain the same across situations. Concepts follow typical patterns of meaning and abstract away from changes. To be sure, there is no contradiction in proposing the concept of a tendency to move away from typical patterns: Change is a constant in human meaning-making, and a concept like “active openness” aims to capture this constant. However, trying to predict the concrete changes by using a strict conceptual apparatus would be a self-contradictory endeavor. A theory cannot foresee what will become relevant to a person here and now. Only another person can do this (and even they will often fail) because human beings, unlike theories, share both tendencies that together make up relevance.

10. Conclusions

I started from communication in a natural language as the prototypical example of a practice involving “signs” in the narrow and popular sense of material phenomena that have or convey meaning. Considering the role and nature of relevance forced me to step beyond this seemingly safe area. The concrete (linguistic) meaning of words in context necessarily involves levels and areas of “meaning” in a wider sense (including. e.g., perception and attention) that was explored, among others, by Schutz. Furthermore, as Schutz pointed out, meaning (Sinn) is not simply something “attached to” a phenomenon. Rather, phenomena are constituted in the first place through meaning [48] (p. 127), and such phenomena as words and other signs are no exception. Meaning in this wider and more basic sense has to do with a selective activity that shapes the way human beings experience themselves and the world around them. This phenomenological analysis appears consistent with that of Peirce, who wondered whether the term “sign” should not be replaced by the more general “mediation” [82]. In addition, Schutz’s view that meaning is a “time problem” and that the selective activity that produces meaning is essentially a process would agree with Peirce’s emphasis on semiosis (rather than simply signs) as the subject of semiotics, on processes—always involving experiencers and their individual and social backgrounds—in which signs are motivated, constituted, interpreted, and modified (ibid.).

Building critically on Schutz, I suggested understanding these processes in terms of a two-sided dynamic of relevance. Conceiving of relevance in terms of an irreducible tension between two tendencies that shape processes of meaning-making appears to account better for human semiotic behavior than do spatializing notions. Typical patterns of meaning give us something similar to a ground we can stand on, but this ground is produced by relevance rather than the reverse, and it is kept in constant motion by the opposite tendency to actively move away from established typical patterns. This latter tendency enables typical patterns of meaning to be truly open by shifting and re-grounding them, but it never gives any guarantees that we have accurately grasped reality or correctly understood each other. Types always remain provisional, and their openness keeps driving on semiosis.

While the active openness of types eludes an idealized rule model or strictly conceptual logic and cannot be predicted by a formal apparatus, it is not a random matter, as I hope to have shown. It involves a flexible and often freely creative, but still public use of typical patterns precisely because such openness is a tendency that human beings share. If this is correct, more research is needed especially to clarify this second tendency, which has mostly been neglected, and to spell out its specific forms and its interaction with the typical grounds that shape the cultural and historical worlds we live in.

Admittedly, the phenomenological “openness” at work here introduces a somewhat perplexing element into a theory of relevance. If a two-sided concept of relevance is applied to semiotics, the perplexing element might also concern and, perhaps, to some extent, unsettle such fundamental concepts as “sign” and “semiosis”. However, this need not be a bad omen. The young Schutz regarded relevance as a (or even the) “central concept of […] the cultural sciences [Geisteswissenschaften]” [83] (p. 3). The mature Schutz added: “Axiomatization and precision regarding the fundamental concept of any science whatever belong to very late stages of its development” [36] (p. 99). “Fundamental” theoretical concepts, too, provide a ground that must be continuously developed, interpreted, and tested through the meaning-making activity of human scientists.