Are Interactive Exhibits at a Science Center Cognitive Artifacts?

The Belgian artist Rene Magritte, with his intriguing paintings, sought to illustrate the claim that “An object never serves the same purpose as either its name or its image does” (Magritte, 2017, p. 35). Compositions such as The Treachery of Images (“Ceci n’est pas une pipe”) highlight the complexity of the seemingly obvious distinction between objects and their representations. Educated adults typically assume that there is a fundamental gap between signs and physical objects—that signs have no causal powers and natural objects are fundamentally non-representational. Magritte, however, showed how fragile this distinction really is.

The problems Magritte pointed out arise also outside the world of art. There are, in fact, many cases when people tend to experience and use representations as if they were physical objects. For instance, one might look at the picture of an apple appreciating its aesthetic value at the same time ignoring its iconic reference to a real or imagined apple. Similarly, a toddler can play with an abacus as a rattle completely disregarding its representational potential. Cases of this sort can be discovered among the various ways visitors at interactive science centers experience the exhibits on offer there. In such cases, the exhibits and installations are experienced in non-representational mode what means that they serve more as an object of direct interest, manipulation or play, than as a representation of particular content or a transmitter of information. This raises the question of whether exhibits at interactive science centers whose representational function has been ignored by the user still fulfill a cognitive function?

The main problem we address in this paper is the semiotic and cognitive status of interactive exhibits at science centers—taking a particular kind of the exhibits (so called open-ended objects) at the Copernicus Science Centre (CSC) in Warsaw as our examples. Intuitively, it would seem that interactive exhibits can unquestionably be classified as examples of cognitive artifacts, that is “physical objects made by humans for the purpose of aiding, enhancing, or improving cognition” (Hutchins, 1999, p. 126). Typical examples of cognitive artefacts include an abacus, a ruler, a written diagram or a map. Counterintuitively, however, some cases of interactive exhibits do not actually fit into the dominant conceptualizations of artifacts performing a cognitive role (Hohol et al., 2021; Kirsh, 2010; Miłkowski, 2022; Norman, 1991; Wartofsky, 1979). As we seek to show, interactive exhibits are only partially compatible with available conceptualizations of cognitive artifacts. This is due to the fact that sometimes representational function of an artifact (interactive exhibit) ceases to be appreciated from the user’s point of view. Is the cognitive function of the artifact also brought into question in this situation? This problem leads to a series of questions discussed throughout the paper. Do existing conceptualizations of cognitive artifacts provide an explanation of the epistemic status of the open-ended interactive exhibits? How might such interactive exhibits be considered cognitive artifacts despite their being experienced in non-representational mode? How can we convincingly describe the process by which a cognitive function of such artifacts is constituted?

In order to tackle these questions we apply alternative conceptualizations of cognitive artifacts which emphasize that a representational relation does not constitute a defining feature of all cognitive artifacts (Heersmink, 2013, 2021; Iliopoulos, 2016; Malafouris, 2013). It is worth noting that classical conceptualization of this concept proposed by Edwin Hutchins remains neutral as far as the representational function of the cognitive artifacts is concerned. Following this lead, we argue herein that an interactive exhibit experienced in non-representational mode can be considered a type of ecological cognitive artifact, as defined by Heersmink (2013) and an enactive sign as conceptualized by Malafouris (2013).

Among the many conceptualizations of cognitive artifacts (Heersmink, 2013; Hutchins, 1999; Norman, 1991; Wartofsky, 1979), it is difficult to find any that takes exhibits in interactive science centers into account. This suggests a certain gap in the theorizing on cognitive artifacts. This paper aims to fill this gap and extend the theory of cognitive artifacts into the area of exhibits at interactive science centers.^{Footnote 1} The question whether interactive exhibits can be considered cognitive artifacts is thus a special case of a more general problem: Can artifacts that are devoid of a representational function be cognitively useful? Our inquiry into interactive science-center exhibits, therefore, seeks to extend and deepen existing analyses of cognitive artifacts. Our aim is to define and introduce an understudied issue of non-representational cognitive artifacts used in interactive science centers. Consequently, the answers provided in the paper are speculative, aiming to outline a potential path towards answering the question whether open-ended exhibits can be considered cognitive artifacts. How far one can go in explaining cognitive role of interactive exhibits without referring to the concept of representation?

The paper consists of three main parts. In Sect. 1 we explain why we apply the theoretical framework of situated cognition. Based on the example of the Copernicus Science Centre (CSC) in Warsaw, we describe the specifics of interactive science centers and characteristic features of interactive exhibits with special attention to so-called “open-ended objects”—as exemplified by the “Swirling Sand” exhibit. Then in Sect. 2, we diagnose the problem of ascribing representational function to interactive exhibits. Then we present standard conceptualizations of cognitive artifacts and explain what obstacles stand in the way of recognizing interactive exhibits as cognitive artifacts. Lastly we demonstrate how interactive exhibits can be understood using the idea of an enactive sign.

1.1 Situated Cognition at the Copernicus Science Centre

In the classical view, knowledge is understood as true and justified belief (Gettier, 1963; Plato, 2014). From this perspective, gaining knowledge means processing abstract contents through reliable research methods in order to produce true beliefs with propositional content. It is traditionally assumed that the data used in the process of knowledge-building is acquired through direct experience or indirectly, through the interpretation of signs. On this view, the external and material objects that people interact with perform merely an instrumental function. They are understood as tools for data acquisition (e.g., a microscope) or as signs transferring encoded content (e.g., a chart or a diagram). Interactive exhibits can play both of these roles. The exhibits show phenomena (e.g., light splitting), and can also show, for example, generalizations of certain mechanisms (e.g., dynamically changing motion diagrams). The thing is, their role is actually much broader. After all, tangible objects such as exhibits are not only tools for information transmission. Their affordances trigger an action with the potential to initiate meaningful sensorimotor experience. The pluralistic model of experiencing exhibits in interactive science centers encourages diverse and alternative modes of use. One of these alternative modes consists in experiencing interactive exhibits in what can be described as a reality mode (Persson, 2008; Sonesson, 2013; Zlatev, 2009). This kind of interaction lies outside the scope of classical theories of knowledge and cognition. Understanding cases in which exhibits are experienced in reality mode requires the application of situated cognition framework.

The central claim of situated cognition framework is that cognitive processes are integrally intertwined with action. Action is not a marginal subsystem that merely provides raw material that is subsequently managed by the cognitive processes proper (i.e., memory or attention). Quite the opposite, action in general, and interaction with artifacts in particular, is an integral part of information processing and knowledge acquisition (Miłkowski et al., 2018; Newen et al., 2018; Robbins & Aydede, 2009; Wilson, 2002). In order to explain the full spectrum of cognitive functions of exhibits at science centers, it is useful to assume that thinking and cognition can be realized through interaction. This tenet is uncontroversial in developmental psychology. Back in the early twentieth century, psychologists showed that children acquire cognitive competence through active and unplanned exploration of the environment. Psychologists such as Piaget have contributed to explaining why bodily interactions with material objects constitute an important dimension of cognition. For Piaget, “knowing an object does not mean copying it—it means acting upon it. It means constructing systems of transformations that can be carried out on or with this object” (Piaget, 1970a, 1970b, p. 15). Similarly, many psychologists have deemed sensorimotor interactions with the environment to be an important, but preparatory stage to the development of abstract thinking (Bruner et al., 1966; Case, 1987; Vygotsky, 1987). Once this stage is reached, sensorimotor operations become superfluous. As children enter a higher stage of development, they become cognitively independent of preceding stages. The visitors of interactive science centers, however, are rarely children at the sensorimotor stage. How, then, can one justify the need for interaction beyond elementary stages of cognitive development? We appeal to the analytical framework of situated cognition because it allows sensorimotor operations to be seen as an important factor in experience, cognition and knowledge acquisition at all stages of cognitive development. In this way we follow the path paved by the researchers who have seen knowledge and cognition as the extension of bodily engagement with the external world (Dewey, 1997; Gibson, 1986; McNeill, 2005).

1.2 What is the Copernicus Science Center?

Science centers are a type of institution of informal education that has flourished in recent years. In Poland, for instance, new centers of this type are being set up every year, and collectively they are being visited by several million people annually. The Copernicus Science Center in Warsaw, one of the largest and most visited science center in Europe, sets itself the mission “to inspire people to experiment, understand the world, and take responsible action.” (Vision, Mission and Values | Copernicus Science Center, 2023). A science center exerts its impact on its visitors via its exhibits: for our purposes the most significant are the tangible objects designed to present natural or technical phenomena that shape the form and scope of a visitor’s experience. The growing popularity of science centers raises the question of the educational value of the exhibitions they offer. How do these objects perform their cognitive function?

From a historical perspective, science centers are the evolutionary successors to private museum collections (such as the Louvre), university collections, and industrial exhibitions (Friedman, 2010). Such exhibition institutions also differ in terms of their conception of the visitor’s role. One such approach can be called “transmissionist”—it assumes that the visitor is the recipient of meanings, defined by the curator in top-down fashion. The second, which can be conventionally called “constructivist,” assumes that exhibitions are meant to stimulate visitors to discover and construct their own meanings. The transmission model of the science museum was formed in the nineteenth century. Back then, science museums were meant to disseminate and popularize knowledge. This concept is rooted in the tradition of the World EXPO; international events aimed at demonstrating the potential of science and technology. Under this model, exhibitions are built around artifacts (e.g. a steam engine, an astronaut’s suit) that symbolize economic development and the achievements of science. Museum exhibitions convey an affirmative message concerning modernization, progress, confidence in science and the role of scientists in society (Iłowiecka-Tańska, & Żabik, 2024). From an anthropological perspective, the nineteenth-century museum tradition relies on a clear separation between the roles of sender (curator) and receiver (visitor). According to this model, the audience is expected to recognize the information that has been previously encoded in the exhibits by the experts. The second, constructivist way of thinking about the cognitive role of science centers has its origins in the mid-twentieth century and its model institution is the Exploratorium—an interactive science center in San Francisco designed by the physicist Frank Oppenheimer, which inspired the creation of hundreds more. Oppenheimer wrote that “A museum can resemble a musical composition, a symphony in which even though the listeners may not be aware of the structure of the piece, they must sense that it exists because the composer was disciplined in his efforts to achieve the coherence of his composition” (Oppenheimer, 1980, p.1). Just as, when listening to a concert, our attention wanders between the different dimensions of the musical piece the visitors in science center wander along their own trail, surrounded by phenomena, more or less consciously choosing which ones they want to stop at. They familiarize themselves with them, driven by their own admiration and curiosity. This metaphor of the museum as a music concert indicates the key features of such a learning center: the importance of the authenticity and personal dimension of the experience.

Under the transmission model of a museum or science center, the relationship between the visitor and the exhibit is clearly defined. It is analogous to the transmission model of communication: just as a recipient is meant to recognize the meaning encoded in a message, a science-center visitor is expected to recognize the information encoded in an exhibit. By contrast, the constructivist model of a museum as adopted by the CSC, assumes that the relationship between exhibits and users is programmatically unspecified and only implicitly suggested. It is subject to creative redefinition, performed each time by visitors. The visitor determines the purpose and scope of the experience she creates. Such a situation raises a number of questions. What is the role of the exhibits in the absence of explicit normative constraints of their interpretation and use? What kind of knowledge or experience do they generate? Do they even perform cognitive functions at all?

1.3 Variety of Interactive Exhibits

The exhibits at the Copernicus Science Center vary significantly in design in terms of form, size, materials, concepts of the interaction. They also vary in terms of their aims: they draw attention to different aspects of a given phenomenon and have different functions—they enable observations or experimentation. Exhibits are often classified into broad categories, e.g. interactive exhibits, hands-on exhibits, open exhibits (Humphrey & Gutwill, 2017). Sometimes designers employ objects visitors will be familiar with, such as a bicycle, a scale, a propeller. Most often, however, exhibits possess unique forms and provide unique possibilities for taking action. Thus a visitor using an object must briefly reflect on what the object is, what is it used for, what possibilities it offers. The more familiar an object is, the less effort it requires to discover its properties and functions. In such a case, the focus on the given phenomenon which is usually the primary goal of interaction appears more quickly and more efficiently, as the perceived affordances of the exhibit suggests what actions to pursue (Chemero, 2003; Dotov et al., 2012; Gibson, 1986). Let’s look at some examples.

The object in Fig. 1 shows an exhibit including a bicycle, an object familiar to visitors. From other contexts, visitors know that it affords a specific action, but in this case, the exhibit becomes a tool through which further exploration of natural phenomena is quickly possible. While sitting on the bike and pedaling, the visitor discovers new functions of the exhibit: their movements are reflected in a mirror that shows the illusion of their skeleton riding a bike right next to them. We call this function-discovery process “instrumentalization.”

Instructions and schematic drawing for the “Pedalling Skeleton” exhibit. An example of perceptual and cultural affordances in exhibit–user interaction

Full size image

Exhibits of another type require the visitor to confront an unfamiliar object, which is unlikely to evoke associations with daily life. Figure 2 shows such an exhibit devoid of familiar cues. Its appearance is unlikely to suggest how it should be used or what the purpose of the proposed interaction might be. It is difficult to grasp the point of the exhibit without instructions or seeing another person interacting with it. In this case, there is usually one correct way to use the exhibit. The procedure to be followed in order to achieve the effect intended by the exhibit’s creators is very particular.

Instructions and schematic drawing for the “Bone Hearing” exhibit. An example of latent affordance

Full size image

Exhibits of a third type, which can be called “open-ended objects,” are designed to be used in a variety of ways (Gutwill, 2008). Such an object aims to create an environment for potentially deep exploration of a phenomenon or process. For instance, the “Swirling Sand” exhibit (Fig. 3) consists of a rotating disc whose speed the visitor can adjust with a knob. The rotating platform is situated in a box of sand. In addition, the exhibit includes tools such as brushes, shovels and trowels. Visitors can use the tools or their bare hands to create patterns in the sand on the disc and study how the sand behaves when it is poured onto different parts of the rotating platform. Just as “Bone hearing” “Swirling sand” exhibit is unfamiliar as a whole. Neither its appearance nor social conventions suggest how it should be used or what the purpose of interacting with it is. However, open ended exhibits do consist of some familiar component parts (e.g., a shovel, a rake, sand) which allows meaningful interaction even in the absence of explicit instructions. Objects such as “Swirling sand”, unlike “Bone hearing” invite immersive engagement even without the recognition of its aims.

Instructions and schematic drawing for the “Swirling Sand” exhibit. An example of open-ended exhibit

Full size image

Such exhibits at the CSC can be said to be of ambiguous status. On the one hand, they are defined by the known and situated context of a science center tour. Visitors know in advance that the activities will be related to science, education, cognition. On the other hand, as the above example shows, certain exhibits possess latent and unknown affordances. Exhibits like “Swirling Sand” have no immediately recognizable function or purpose. This tension between known context and unknown objects is what inspired us to consider the role of exhibits conceived as cognitive artifacts.

1.4 Double Semiotic Link at School and in CSC

There is a rich body of research on the artifacts, tools and manipulatives used in school education (Bruner, 1974; Piaget, 1970b; Vygotsky, 1987). It is worth explaining what the difference is between research focused on school education and the specificity of the problem addressed in our paper. We will use a review paper by Bussi and Mariotti (2008) as an example to show the difference. The authors analyze the use of artifacts such as the compass and the abacus in the classroom.^{Footnote 2} Bussi and Mariotti draw a description of a didactic cycle where students are encouraged to perform activities with the artifacts, then they are asked by the teacher to individually produce some drawing or written record. Finally, students engage in a collective exchange of narratives, oral utterances or texts about their actions related to the task (Bussi & Mariotti, 2008). There are at least three important features common to all of these activities. First, students are monitored and supervised by the teacher, who delivers cues and directs their attention. Second, their tasks are explicitly described and well-defined. Third, the class environment is highly and explicitly normatively constrained. For example, pause is clearly separated from the lesson, and students are well aware of when to speak and when to remain silent.

Bussi and Mariotti (2008), build on Rabardel’s and Vygotsky's ideas and hypothesize that the characteristic feature of artifacts used in school settings is their double semiotic link (DSL). The link is created on the one hand between artifact and task and on the other hand between artifact use and abstract knowledge.^{Footnote 3} As the student confronts the task with the use of an artifact, she starts recognizing that the artifact possesses certain possibilities and constraints for solving the task. Using the artifact becomes a meaningful activity linked to obtaining some aim. In this way, the first dimension of the double semiotic link (DSL) is created. It draws together personal meanings with the use of the artifact to perform a task. The use of artifacts becomes a meaningful activity directed toward fulfilling some aim. The second dimension of the DSL requires more explicit instruction on the part of the teacher, who helps students link the use of the artifact with abstract concepts. According to Bussi and Mariotti (2009), creating the link between artifact use and abstract concepts (the second dimension of the DSL) is essential for successful education. Teachers are supposed to guide the students to make a gradual transition from meaningful activity with artifacts (the first dimension of the DSL) to grasping that their activity is related to abstract knowledge (the second dimension of the DSL). They stress that in order to describe the double semiotic link (DSL):

We do not refer only to the concrete act of using a tool to accomplish a task, rather to the fact that new meanings, related to the actual use of a tool, may be generated, and evolve, under the guidance of the expert (Bussi & Mariotti, p. 754).

The situation in the classroom is significantly different from that in the CSC. Visitors to the interactive museums are invited to spontaneously interact with the exhibit. Unlike students in the classroom, they have no guidance from an educator and can freely undertake activities on their own. Usually, visitors have no tasks at all or their tasks are ill-defined, without one correct method or solution. Although there are implicit norms inscribed into the context of the visit in the educational institution, CSC programmatically encourages visitors to do whatever they please with the exhibits on offer. These differences determine that the model of the double semiotic link (DSL) does not fully translate into the analysis of exhibits at the CSC. Since there is no expert who guides visitors' actions and attention, it is difficult to speak about the second dimension of the DSL (the relationship between meaningful action and abstract content). As Bussi and Mariotti (2008) spell it out, in order to create the second dimension of the DSL “the use of the artifact has to be fully integrated in the classroom activities (…). The role of the teacher is then crucial” (p. 754).

The first dimension of the DSL is also problematic, since visitors usually have no clearly specified task to be solved. There could hardly be a relationship established between personal meaning and the use of artifacts for a certain aim. It does not mean that visitors’ activities are aimless, but their aims are pretty vague as compared to the classroom environment. An additional problem is that exhibits in CSC can be treated by the visitors either as referential signs or as objects in their own right. Open-ended exhibits described in the previous section have relevant features that reduce the likelihood of treating them as a referential signs by the visitors. They are very attractive in their own right and invite engaging interactions. These characteristics increase the possibility that visitors treat them as objects of direct interest and play, rather than signs referring to some content (DeLoache et al., 2003; McNeil et al., 2009; Uttal et al., 1997). Additionally, the internal CSC surveys indicate that only about half of the visitors read the instructions accompanying the exhibits, and even fewer read them with understanding. Open-ended exhibits, in particular, afford intuitive and self-guided interaction that is why they are frequently used without consulting the information boards (Krejtz, 2017; Mieszczanek & Elbanowski, 2013).^{Footnote 4} This suggests that visitors to CSC certainly display meaningful activities related to the exhibits, but they often do not use the open-ended exhibits as representations. How to spell out the cognitive potential of open-ended exhibits, taking into account all the specifics of CSC settings?

If visitors at the CSC are interested in open-ended objects as things themselves rather than representations, this might call into question the cognitive functions of the exhibits. This leads to the further question of whether interactive exhibits that resemble “open-ended objects” can rightfully be classified as cognitive artifacts—material objects used to transmit, process, or store information.

2.1 Meaningful Objects and Representational Artifacts

In order to understand what it means that some exhibits at a science center may not be used as representational sign, we need to clearly distinguish the concept of representational function from the concept of meaning. We apply a broad definition of meaning, as a relationship between an organism and the environment determined by the interests and values of that organism (Sonesson, 2006; Zlatev, 2000, 2009). Meaning is displayed by a broader class of semiotic objects and processes than the representational function is. An object can be meaningful even without a representational function. In this case, any object that an agent focuses her attention on or directs her action toward is a meaningful entity. It is simply distinguished from the environmental background according to some idiosyncratic aim or value. Meaningful objects of this kind appear at the very basic biological level. A classic example is the odor of butyric acid, conceived as a meaningful signal for a tick (von Uexküll, 1957). At a higher level of biosemiotic phenomena, one can also talk about affective meaning. This requires minimal phenomenal consciousness, which a tick does not possess (Sonesson, 2006, 2012).

Not all meaningful objects are representational. An object that is both meaningful and representational is usually called a sign or a symbol (Peirce, 1992; Piaget, 2005; Piaget & Inhelder, 1969). Developmental psychologists have recognized that discovering a representational function is not an easy task for children. In order to identify an object as a sign, not merely a meaningful entity, they must be able to distinguish between (at least) two semiotic dimensions within it: namely, between the expression (the sign vehicle) and its content (the object of its reference or mental picture associated with it). Without discriminating between what the sign is as a physical object from what it points to (another physical, mental, abstract object), it is impossible to appreciate its representational function. Let’s take a urinal as an example. It is a meaningful object which is normally used for practical aims, but in certain circumstances, for instance at museum’s exhibition, a urinal can take on a representational function. As long as a urinal is in a public restroom its instrumental value is primarily appreciated. It is undoubtedly meaningful, but it is not a sign since the user does not normally distinguish between its two semiotic dimensions: expression and content. However, when Marcel Duchamp moved a urinal into an art gallery in 1917, these two semiotic dimensions were immediately distinguished. The urinal obtained a representational function. A representational function is a relational feature. It is not an inherent property of an artifact. Representational function emerges out of the interaction between the agent and the external world and the social conventions.

2.2 Reality and Pictorial Mode of Experience

Discovering a representational function is sometimes problematic even in cases of objects to which educated adults ascribe a representational function by default, such as a realistic photograph. Studying the semiotic sensitivity of great apes, Persson (2008) describes three modes of image perception: the surface mode, reality mode, and image mode. Depending on the subject’s conceptual resources and the context of interaction, any image (e.g., a photographic picture of an apple) can be experienced on each of these levels. Depending on the subject’s experience and abilities, she might perceive a picture of an apple in surface mode. She perceives colors, shapes and patterns, but she does not recognize that the photo shows a meaningful object, namely an apple. In the reality mode of picture perception, the subject recognizes the figure of an object (e.g., an apple) but confuses the picture with reality. This happens, for instance, when a subject perceives a flat, two-dimensional figure of an apple and tries to reach out for it, as if she were dealing with the object of its reference. The fact that a subject (e.g., a child or great ape) treats a two-dimensional photographic picture like a three-dimensional object indicates that she is confusing expression and content. The picture is experienced as an object for instrumental action, rather than an object of interpretation. In this special case an agent acts upon a picture as if it were a real object, not as an object designed to refer her to some aspects of the world (Persson, 2008, pp. 7–15, 17–24, 43–53; Zlatev et al., 2013).

In both the surface and reality mode of picture perception, the user fails to appreciate the semiotic duality of the image. From the user’s perspective, the expression and the content are so integrated that they form an ordinary object, similar to other objects in the environment. When experiencing an image in reality mode a person does not recognize its representational function. Spectacular instances of these phenomena typically occur in art galleries, when cleaning crews attempt to remove or wipe off elements of an exhibit. For example, in 1986, cleaning staff at the Dusseldorf Academy of Fine Arts failed to distinguish between expression and content when they decided to wash Joseph Beuys’ grease-stained installation “No Title (Bathtub).”

The representational function appears when an artifact (e.g., a work of art or a photo of an apple) ceases to be considered as an actual object, but instead as an object that refers to something other than itself (represent a real or imagined apple). This happens in the pictorial mode of experience (Persson, 2008). Despite possible confusion between reality and representation (DeLoache et al., 2004; Olson, 1994; Trybulec, 2017) educated adults usually have no problems in recognizing when expression and reference should be separated.

The representational function bears at least three formal characteristics. First, from the user’s point of view expression and content are temporally and spatially divisible (Sonesson, 2007, 2013). If an object is experienced as a sign, its expression is supposed to point beyond itself, toward something that is different than the mere expression. It follows that in the user’s experience expression and content appear as qualitatively different phenomena. That is why in standard circumstances the user does not confuse a sign with its reference. Second, the relationship between expression and content is asymmetrical. While it makes sense to say that a photograph represents the actual apple depicted, it is unreasonable to claim that the actual apple represents the photo (Costall, 1997; Goodman, 1968). Third, asymmetry between expression and content can also be observed at the level of the user’s experience. On the one hand, although the sign user directly perceives the expression (sign vehicle), she does not direct her conscious attention toward the expression, but instead focuses her attention on the content of the sign. On the other hand, the content of the sign (i.e. the actual or imagined apple), although not directly experienced, is consciously attended to. This is the case in the pictorial mode of picture perception. In reality mode, the user’s experience is quite different. The user not only fails to distinguish between expression (sign vehicle) and content (reference), she moreover fails to experience the asymmetrical relation between the two (Persson, 2008; Sonesson, 2012, 2013). Therefore, in the reality mode of picture perception, the user does not appreciate its representational dimension.

The reality mode of experience is of particular importance to our argument, because interactive exhibits at a science center like the CSC are often experienced by young visitors with a sort of neglect for the exhibits’ representational function. For instance, when the “Swirling Sand” exhibit (see Sect. 1.2) is used as a kind of a sandbox, young visitors appreciate its entertaining value and disregard what the exhibit is about from the perspective of its designer. This suggests that they tend to ignore the difference between two semiotic dimensions implicitly assumed in the exhibit: expression and content. In cases of experiencing exhibits in reality mode, the representational relation evaporates. And yet, standard conceptualizations of cognitive artifacts stress that the representational function is a necessary condition for the cognitive role to be fulfilled (see Sect. 3.1). Therefore the question emerges as to whether exhibits experienced in reality mode actually have any cognitive role.

2.3 Two Ways of Recognizing the Representational Function

Recognizing the representational function requires a complex semiotic consciousness on the part of the user, who is meant to be capable of grasping the difference between expression and content. In order to recognize the difference between a meaningful object and a sign, one needs to apply a first-person perspective. In principle however, the problem of the representational status of interactive science-center exhibits can be analyzed from either a first- or a third-person perspective (Konderak, 2018, pp. 28–30, 111–127; Sonesson, 2012; Zlatev, 2012).

The description of an exhibit from the third-person perspective tends to deliver all objectively observed data about the artifact as well as whole situation and the circumstances where it is situated. This kind of description also concerns the institutional context of the exhibition (museum or science center) and normatively structured and standardized interactions. In a broad sense, even the most interactive exhibits, as elements of the science-center institutional context, refer to implicit and normatively constrained ways of dealing with them. Moreover, characterization of exhibits from the third-person point of view involves the curator’s idea of the artifact as well as the intentions of its creator, usually available in a verbal description accompanying the exhibit. For example, according to the description presented at the CSC next to the “Swirling Sand” exhibit (see Fig. 3, Sect. 1.2), this artifact is meant to inform users in an accessible manner about the physical principle of relativity of motion. By interacting with the exhibit, the visitors do not need to imagine what the relativity principle is about. Rather, they can physically design and enact a situation in which the relativity principle is visible and can be felt through their bodily experience (e.g., a stationary finger on a moving platform). From this perspective, exhibits undoubtedly display a representational function.

A different picture emerges, however, when the perspective of the user is taken into account. The visitor’s first-person experience indicates that the representational function of exhibits is often neglected. It is a fact that various visitors interact with exhibits in various ways (Potęga vel Żabik et al., 2024). During exploration, some of them do discover the so-called “intended effect” of the exhibit; the phenomenon that the creators intended as the main purpose of the exploration. However, a significant group of visitors ignore the description of the exhibit and focuses on the exhibit’s properties as if it were a kind of a toy or a machine to play with (Gutwill, 2008). This is often the case particularly with open-ended exhibits, such as “Swirling Sand” (Fig. 3). From this perspective, an exhibit simply is an object on its own right. As the object of visitor’s direct interest, it does not refer to anything beyond itself. In this case, visitors experience the exhibit in reality mode. This is analogous to the situation with photographs (see Sect. 2.2). Just as in perceiving pictures in reality mode the object of interest is the picture itself, in the case of reality-mode experiencing of interactive exhibits, the object of interest is the exhibit itself (DeLoache et al., 2004; Uttal et al., 1997). In this situation, talking about the representational function of the artifact becomes problematic. The visitor interacts spontaneously with the exhibit and does not distinguish between its semiotic dimensions: expression and content. Her direct experience and conscious attention are both directed at the exhibit itself. From the user’s perspective there is no asymmetry between the expression (the sign vehicle) and its content. What, then, is the cognitive role of the interactive exhibits if users do not recognize their representational function? Does an interactive exhibit become a mere toy, merely providing fleeting entertainment? Or is it possible that even when experienced in reality mode interactive exhibits can nevertheless perform a cognitive role? If this is the case, how are such interactive exhibits able to become cognitive artifacts?

Can open-ended exhibits experienced in reality mode be considered cognitive artifacts? Are such interactive exhibits actually tools that support the acquisition, storage, and processing of knowledge and information? To address this questions, we will briefly review existing classifications and standard conceptualizations of cognitive artifacts. These provide detailed criteria to assess when objects or tools perform cognitive functions. If interactive exhibits possess all the relevant features of a cognitive artifact, then their cognitive function should not be questioned.

3.1 Classifications and Definitions of Cognitive Artifacts

Cognitive artifacts are commonly defined as material devices that perform representational function and enable the user to store, transmit, and process information (Heersmink, 2013, 2021; Norman, 1991). Canonical examples of cognitive artifacts include diagrams, notes, sketches, maps, and calendars. In the literature, cognitive artifacts are classified according to various criteria. Dascal (2002) proposes to classify cognitive artifacts (which he dubbed cognitive tools) according to five criteria, resulting in five pairs of opposing categories. Table 1 presents his classification in a simplified form.

Dascal (2002) distinguishes strong and weak cognitive tools on the basis of the modal properties of beliefs which emerge out of the application of the tool. A mathematical proof is a strong cognitive device because from true premises it always leads to true conclusions. In contrast, engaging in a discussion often produces rather weak (inferential) outcomes. By the second criterion, a computer provides an instance of an integral cognitive tool, since it requires minimal activity on the part of the user, i.e. the human needs to merely deliver input data and read out the output. In contrast, devices such as an abacus are partial cognitive artifacts, since they involve continuous activity on the part of the user during information processing. The third criterion distinguishes between artifacts that comprehensively or partially describe their domain of reference. For example, a sign system for traffic fully describes its domain of reference, unlike, say, the rules of literary narrative, which touch upon certain aspects of the described reality (such as the psyche of the protagonist). According to the fourth criterion, an artifact is constitutive if it is necessarily required to achieve the desired cognitive results.^{Footnote 5} Symbolic representations of numbers are considered constitutive, since they allow humans to perform exact computations on large quantities (Dehaene, 1999; Hohol et al., 2021). The final criterion concerns the location of the cognitive artifact. A shopping list is an external artifact, whereas mnemonic mental strategies are basically internal (Dascal, 2002).

Wilson and Clark (2009) provide a simpler classification (Table 2), considering only two criteria. Based on origins, a distinction is drawn between natural, technological, and socio-cultural tools. These include means as basic as odor cues, but also diagrams, notes, measuring devices, writing systems, division of labor and educational practices. The second criterion, in turn, pertains to the temporal characteristics of cognitive systems consisted by humans and their cognitive tools. A book left on a desk might be a spontaneously created mnemonic cue, suggesting that someone should read. This kind of cue provides a temporary cognitive tool that is situational and idiosyncratic. In contrast, natural language is one of the most powerful and commonly used cognitive artifact (Carruthers, 2008; Clark, 2010; Wilson & Clark, 2009).

These classifications of cognitive artifacts provide a conceptual framework which can help us to identify the essential features of interactive exhibits.^{Footnote 6} One might classify interactive exhibits such as “Swirling Sand” as an external, weak, partial, incomplete and non-constitutive, technological, and repetitive cognitive artifact. It is worth highlighting that among the many objects cited in the literature as examples of cognitive artifacts, there are none that directly refer to a museum’s or science centers’ exhibits. This may suggest a certain gap in the theorizing on cognitive artifacts. Most importantly, however, the classifications utilize categories and examples of cognitive artifacts that display representational function. These are objects that refer to something other than themselves. The classifications suggest that a cognitive artifact is an object that performs a representational function.

Moreover, the conceptualizations and definitions of cognitive artifacts enhance the tendency observed in the classifications to frame cognitive artifacts in terms of their representational function. An example is provided by Marx W. Wartofsky, who proposed one of the earliest conceptualizations of cognitive artifacts. He writes that.

we create cognitive artifacts which (…) radically alter the very nature of learning and which demarcate human knowledge from animal intelligence. The cognitive artifacts we create are models: representations to ourselves of what we do, of what we want, and of what we hope for (Wartofsky, 1979, p. XV) [emphasis added].

Commonly used definitions reaffirm Wartofsy’s stance. Donald Norman, in turn, stresses that a “cognitive artifact is an artificial device designed to maintain, display, or operate upon information in order to serve a representational function” (Norman, 1991) [emphasis added]. Other researchers replicate the conceptual schemes present in Wartofsky and Norman (Fasoli, 2018; Nersessian, 2005; Tweney, 2002).

Interactive exhibits such as “Swirling Sand” certainly meet some of the defining characteristics of a cognitive artifact, so construed. It is a material object intentionally produced by people in order to store, transfer and manipulate information. Interactive exhibits have no immediate practical purpose; therefore, they are easily distinguished from technical artifacts. It is not evident, however, whether interactive exhibits are closer to cognitive artifacts than to artifacts that provide mere entertainment. In order to resolve these concerns, one might wish to invoke the distinctive feature of cognitive artifacts: their representational function. The connection between representational and cognitive function of an artifact stems from the assumption that for an object to have a cognitive role it must be “about something” (represent something). If an object is “about something” it can provide some information about reality and thus it can expand the user’s knowledge. However in the Sects. 1.3 and 2.2 we have explained that the representational function of interactive exhibits remains problematic. Hence, if interactive exhibits do not perform a representational function, under the standard view on cognitive artifacts their cognitive value becomes questionable.

3.2 Interactive Exhibits as an Enactive Signs

The discussion so far has shown that open-ended exhibits experienced in reality mode, due to their problematic semiotic status, do not fit into standard theories of cognitive artifacts. In particular, from the user’s perspective some of the exhibits do not perform a representational function while most definitions and classifications of cognitive artifacts particularly emphasize their representational character (see Sect. 3.1). It can be argued, however, that standard conceptualizations place undue emphasis on sign-like artifacts (Heersmink, 2012, 2013; Malafouris, 2008, 2013, pp. 137–143). Following this path, Richard Heersmink distinguishes a class of ecological artifacts whose cognitive utility is not determined by the semiotic relation between the sign vehicle (expression) and the content, but by the pragmatic relationship between the user and the object. Interactive exhibits can be framed as a special type of ecological cognitive artifacts singled out by Heersmink (2013, 2021).^{Footnote 7}

How might the cognitive role of open-ended exhibits at the CSC be explained without referring to a representational function? According to Heersmink (2013, 2021), the representational function is not a necessary condition for an artifact to perform a cognitive role. While the class of ecological cognitive artifacts identified by Heersmink encompasses interactions with open-ended exhibits, it does not provide a detailed explanation of the mechanism of their cognitive impact. To address this issue, we draw on the concept of the enactive sign developed by Lambros Malafouris, which exhibits conceptual similarities to the idea of ecological cognitive artifacts (Iliopoulos, 2016; Malafouris, 2007, 2008, 2013). Both Malafouris and Heersmink aim to explain how material objects can support cognitive processes without relying on representational function. However, as an archaeologist, Malafouris proposes a broader analysis of artifacts across time and space, exploring how everyday objects and practices, such as body ornaments, pottery making and stone tools, contributed to cognitive development at the dawn of human culture (Malafouris, 2008, 2013, 2018). The idea of enactive sign offers broader scope and deeper depth of explanation than the concept of ecological CA.

Malafouris applies a broad concept of a sign understood as a meaningful object. He introduces an important distinction between a referential sign and an enactive/material sign. In the terms applied in this paper, his distinction corresponds to the difference between a sign and a meaningful object without a representational function (see Sect. 2.1). Malfouris argues that “the material sign is constituted as a meaningful entity not for what it represents but for what it brings forth: the possibility of meaningful engagement” (Malafouris, 2013, pp. 104–105) [emphasis added]. Although the material or enactive sign does not perform a representational role, it nevertheless has an enactive function.

What is an enactive function? It means that an enactive sign can evoke and create a meaningful experience.^{Footnote 8} It does not do so by reference to phenomena or concepts that exist independently of the enactive sign. Through the enactive function the sign embodies these phenomena or concepts. The material sign does not stand for phenomena that are something other than itself, but itself makes them present and reifies them. For Malafouris, a referential function presupposes that it expresses content that existed before the sign was created or used. Therefore a referential sign points to some reality that is temporally and ontologically prior to the sign. The function of reference is passive, since it is not actively involved in the creation of its referent. The reference function does not co-create the object of reference; it indifferently describes the reference. An enactive sign, by contrast, has performative functions. Unlike a referential sign, it is a dynamic force that shapes reality and co-produces mental phenomena. By virtue of its dynamic and performative character, an enactive sign does not need to be recognized by the user as an object with a referential value, yet it can nevertheless support cognitive processes (Malafouris, 2008, 2013, pp. 89–99, 144–148, 164–177).

By what mechanism can an enactive sign perform a cognitive role? The key is the process of integrating various domains of the user’s experience by virtue of creating associative links between them. The process of integration supports direct connections and promotes similarities between very distant domains of an individual’s experience. This process allows the features of directly observed objects (e.g., their spatial properties) to be unconsciously projected into abstract concepts and relations that are sometimes vaguely understood (Alač, 2003; Alač & Hutchins, 2004; Becvar et al., 2005).

An object as ordinary as a queue in a grocery store provides a good example of such a projection (Hutchins, 2005, 2014). Described from a third-person perspective, the queue certainly has a representational character. It refers back to the history of the customers’ appearance. However, from the first-person perspective of people involved, the queue is primarily a tool for guiding actions and decisions regarding the order of service. It is more a tool to organize social relations and to impose social order, rather than a sign to describe it. Imperceptible and difficult-to-remember temporal relations are reified in the form of the stable and material structure of a line. People understand the line as a queue through replacing the abstract temporal order by concrete spatial relations between them. In normal circumstances the customers have no reason to mentally represent the temporal order of appearance of other people in a queue, because the queue is present. In Sonesson’s (2007) words one can say that there is no asymmetry between attention and perception, in this case. Customers monitor who is standing where and direct both their attention and perception to the queue as a material object. As long as no one challenges the staus quo, they do not need to refer to abstract temporal relations. In this sense, the queue is an enactive sign. It does not refer beyond itself, but embodies and enacts abstract temporal relations. As an enactive sign, the queue “can be seen to operate simultaneously as a signifier and signified (…) both possibilities are equally afforded” (Malafouris, 2013, 117) [emphasis added]. From the first-person perspective, the queue is undoubtedly a meaningful object, but not by virtue of referring to something beyond itself. For the people involved, the queue does not represent the sequence of customers’ appearance, it enacts the sequence of service. The real function of the queue is that the spatial order of people performs a social order and in this way determines the temporal order of service. As an enactment of the eligible temporal sequence of service, it gains a normative dimension by performing the order of acting and thinking (Hutchins, 2005, 2014; Malafouris, 2013, pp. 90–118).

Interactive exhibits experienced in reality mode can be described much like a queue in a grocery store. From a third-person perspective, an exhibit such as “Swirling Sand” undoubtedly fulfills a representational role (see Sect. 2.3). However, the first-person experience of the young visitors suggests that they do not always recognize a representational function of the exhibit. Are they, then, using the exhibit incorrectly in such cases? As argued in Sects. 1.1 and 1.2 above, interactive science centers such as the CSC put strong emphasis on visitors’ unconstrained experimentation and encourages them to freely explore the exhibits. Therefore, the non-representational use of an interactive exhibit is considered just as appropriate as the other modes of use.

Does “Swirling Sand” display a cognitive role when used in non-representational mode? The idea of the enactive sign helps to spell out cognitive value of interactive exhibits used in reality mode. For instance, the cognitive role of the “Swirling Sand” exhibit does not necessarily consist in representing the principle of relativity of motion, but in enacting the relativity of motion. When a visitor places her finger on the rotating platform, it allows her not only to imagine what the relativity of motion might be, but also to experience it bodily. The visitor feels, not only sees, that her finger is stationary. She also feels that the sand scattered on the rotating platform is also stationary in relation to the platform but not in relation to her still finger. She feels through her body what the difference is between a stationary finger and stationary sand. She directly experiences what a frame of reference is, since she directly senses her own body as a frame of reference. The “Swirling Sand” exhibit used in reality mode allows her to experience the relativity of motion through embodied engagement, a different modality than mere visual perception or mental imagination conventionally construed. The actual bodily experience of playing with “Swirling Sand” consists in enacting the principle of relative motion, even when the visitor has no explicit knowledge about this principle. This point can be illustrated by the research on the role of gestures during school education. Goldin-Meadow (2015) explains that it is often the case that students who are expected to master new and abstract concepts unconsciously express them with gestures before expressing them in speech. For instance, Goldin-Meadow observed that some children solving a Piget’s conservation task provide a false answer while at the same time making a specific gestures, suggesting they begin to grasp the idea of substance constancy. Particularly interesting is the group of children who display disparity between incorrect verbal answers and “correct” gestures, suggesting a partial understanding of the problem. They are in the transition between incorrect hunches expressed in words and correct notion implicit in gestures. Individuals who are in this transitional stage are more sensitive to instructional cues from the teacher. This suggests that sensorimotor engagement supports the development of an implicit and tentative grasp of abstract concepts. One of the practical conclusions from Goldin-Meadow studies is that the teachers can draw on sensorimotor experience in order to help students find their way to explicit understanding (Beilock & Goldin-Meadow, 2010; Goldin-Meadow, 2015).

The physical structure of the open-ended exhibits encourages the user to manipulate, touch, and transform it. It allows for meaningful engagement by physically performing the relative motion using the visitor’s body tuned to the movement of the platforms. Open-ended objects work as facilitators or enablers creating sensorimotor engagement that can help visitors to grasp abstract concept. Such activity creates a possibility to integrate particular bodily experience with the abstract notion of relative motion and frame of reference later on in education in a way similar to that of gestures in McNeil’s and Goldin-Meadow’s research. Gestures as well as actions performed on open-ended objects, do not display a representational function, but prepare the student to master abstract content (Beilock & Goldin-Meadow, 2010; Goldin-Meadow, 2015; McNeill, 2005). By manipulating an exhibit such as “Swirling Sand,” the visitor simultaneously enacts relative motion, in the same sense as children in Golding-Meadow studies enacts substance constancy even when incorrectly verbally responding.^{Footnote 9} Engaging in such meaningful interactions can, over time, serve as the raw material that will be used to master currently inaccessible and abstract content.

Even if interactive exhibits are used in reality mode they can perform a cognitive function. In Malafouris’ terms, interactive exhibits are a special case of enactive signs, which enable abstract principles or concepts to be enacted by the visitor’s bodily experience. Physical interaction with the exhibit transforms incomprehensible abstract content into concrete experience, which can be used in subsequent stages of education. The process of enactive signification involves integration of intangible and abstract concepts with a concrete bodily experience, gained by exploring material artifacts. In this way it supports the claim that interaction with an enactive sign can indeed have cognitive functions, despite lacking a representational dimension.

Interactive science centers such as the Copernicus Science Center in Warsaw (CSC) house thousands of interactive exhibits. According to the manifesto of the CSC, visitors’ interactions with such exhibits are organized in a bottom-up fashion: the designers of the exhibits do not impose any particular, “correct” model for experiencing them. Quite the contrary, visitors are encouraged to explore the exhibits spontaneously, according to their own needs, goals and abilities. This results in a great variety of ways of experiencing the exhibits. One significant way by which visitors explore the exhibits involves using them in reality mode (Persson, 2008; Zlatev, 2009). In this form of interaction, the exhibit is put to work not as a representation conveying specific content, but as an object of direct interest, play and entertainment. This form of interaction is particularly common with certain exhibits of the open-ended type (such as “Swirling Sands”). From the user’s point of view, in reality mode, the referential relation between expression (the sign vehicle) and content (the object of its reference) of the exhibit disappears. Can such exhibits, nevertheless, perform cognitive functions in such circumstances?

We applied the concept of cognitive artifact, seeking to identify whether it can be used to make sense of the interactive exhibits’ cognitive functions when experienced in reality mode. We showed that standard conceptualizations of cognitive artifacts (Dascal, 2002; Fasoli, 2018; Greif, 2019; Norman, 1991; Wartofsky, 1979) are unsuitable to explain cognitive functions of interactive exhibits, because they exclude objects without representational function. We argued that representational function does not need to be defining feature of being a cognitive artifact, it is an important dimension of many cognitive artifacts though (e.g., writing, abacus etc.). (Heersmink, 2013, 2021; Iliopoulos, 2016; Malafouris, 2013). Then we argued in favor of the claim that interactive exhibits used in reality mode can be considered a special case of nonrepresentational ecological cognitive artifacts in Heersmink’s sense (2013, 2021). By invoking Malafouris’ concept of enactive sign (Iliopoulos, 2016; Malafouris, 2008, 2013) we explained how the mechanism of their cognitive impact can be understood. The cognitive functions of non-representational artifacts is performed by integrating different areas of experience. Interactive exhibits used in reality mode do not represent, but rather make present or perform the phenomena and regularities that are the object of informal education in CSC. In this way, they are used to enact and evoke a meaningful experience that forms the basis for further education and knowledge acquisition outside the interactive science centers.