Gordon Pask’s second-order cybernetics and Lev Vygotsky’s cultural historical theory: Understanding the role of the internet in developing human thinking

Pask’s foretelling of neoclassical cybernetics

At Cambridge, after reading Norbert Wiener’s (1961) Cybernetics and the central tenet of humans and machines processing information through feedback loops to retain stability, Pask took a deeper interest in cybernetic theory. In 1959 his attendance at the conference “Mechanization of Thought Processes” at the National Physics Laboratory in London marked his formal entry into the field. Pask’s paper, “Physical Analogues to the Growth of a Concept” (Pask, 1959a), provided theoretical explanations of learning and concept development accompanied by physical demonstrations representing fulfillment of a contingency or reinforcement of skill (e.g., growth of crystals in electrolytic solution; Scott, 2007). He had begun building responsive devices and founded an organization named System Research Ltd (Pickering, 2010). Upon deciding to study cybernetics, Pask displayed understanding of the field by building tools that created dynamic light shows for musical performances (Pickering, 2010). After meeting Wiener, Pask worked with operations research scientist Stafford Beer to understand whether biological processes could be formalized in baroque machines. Ashby’s (1952/2016) homeostat was central to their investigations. The machine comprised four electromagnetic units connected by dynamic feedback loops, representing organisms interacting with the environment. Input from manipulating vanes would disturb every other unit, until each vane reached stability.

Beer and Pask wanted to use living organisms as prototypes for Ashby’s synthetic brains. Pask investigated possibilities to use mosquito larvae, while Beer investigated freshwater crustaceans and pond ecosystems. They were unable to find organisms whose physical/biological essence could be replicated to resemble the homeostat’s functioning (Pickering, 2009). They achieved some success through “fungoid whisker systems” comprising three platinum electrodes in ferrous sulphate solution. Pask and Beer dangled the devices on a window in Pask’s apartment on Baker Street in London, training them to discern segments on a noise spectrum, with the device effectively “evolving an ear” (Beer, 2001). Pask conceptualized “survival mechanisms” as rewards for the machine, repaying correct discernment of sound with increased current. Even when wires were severed, applied current regrew connections between anode and cathode.

Beer and Pask’s chemical computers could be reinforced to respond to certain stimuli, but this conceptualization only represented a precognitive mind capable of finite outputs, like Ashby’s (1952/2016) homeostat. The two cyberneticians further diversified their work to understand how human agency and social contexts (with limitless possibilities) could be incorporated into cybernetics (Pickering, 2010). Pask focused on education and performance art to attune computing machines to social processes.

Pask and Beer later joined polyglot cybernetician and electrical engineer Heinz von Foerster at the Biological Computer Laboratory (BCL) at the University of Illinois, Urbana as visiting scholars. Commenting on Pask’s work, von Foerster suggested he inserted the idea of the “game” into the BCL narrative, considering knowledge development as an active process involving interactions and understandings of phenomenological qualities of language and experience, rather than mechanistic input–output scenarios (von Foerster, 1993). Attending the “First International Conference on Self-Referential Systems” organized by David Wells, Pask extended biologist Humberto Maturana’s (another BCL scholar) ideas about the organizationally closed brain–body system. Pask discussed the circularity of psychosocial conversations, and suggested biomechanical systems are taciturn, reaching varying stability by changing internal states responding to environmental stimuli (Scott, 2004). Conversational systems interact, self-producing new meanings as observers of activity in context. In his paper, “The Natural History of Networks,” Pask (1959b) explains that self-organizing systems react to the environment to constantly reconstruct their cultural reference frames (Scott, 2021). By outlining distinctions and unities between brain–body and psychosocial systems, Pask foretold emergence of a neoclassical cybernetics; what von Foerster later called second-order cybernetics.

Applying cybernetics to education during the 1960s and 70s, Pask developed interfaces supporting learning by spurring human–human, human–machine, and machine–machine conversations. He employed conversation theory, which expanded the distinction of biomechanical and psychosocial systems using the nomenclature of M-individuals and P-individuals, to understand how observers reconstruct concepts, forming semantic web-like structures in the mind through shared activity, responding to incoming information (Pask, 1976). In Pask’s collaborative classroom, each learner occupies novice and expert positions, producing lateral, nonhierarchical communication with peers, instructors, and interfaces to understand the how and why of knowledge (Scott, 2001). Collaboration fuels learning and concept development. Machines may assist collaboration, becoming surrogate participants in learning conversations and mediating human-to-human interactions.

Pask’s (1975, 1976) framework of collaborative learning heavily draws from Vygotsky’s (1997) ideas on relationships between action and thinking, suggesting that the development of mental models through speech and abstraction begin as activity in the environment. He explicitly mentions Vygotsky in explaining how language-oriented systems or psychosocial systems are always embedded within social realities (Pask, 1976), morphing mental models and actions used to embody activity. In the next section, we outline links between the work of these two scholars, and show how conversation theory may provide an information age mirror to Vygotsky’s approach.

Integrating Pask and Vygotsky

Conversation theory suggests understanding human thinking and behavior requires decoding behaviors that reformulate “concealed” mental phenomena (Pask, 1976, pp. 19–20). Pask (1975, p. 71) says language or L in a conversation can be spoken or written for living systems, but may be conceptualized as graphical displays and interactive prompts when considering technology, allowing both pragmatic and semantic interpretations towards reaching shared understandings (Pask, 1975, p. 22). Much like conversational analysis approaches propounded by Sacks et al. (1974), Pask (1975, pp. 165–166) says the action of conversing itself, rather than mental models, may have time and place. However, rather than using turn-taking approaches to only understand actions in conversations within closed settings, Pask aims at looking at how machine and human-initiated conversational feedback loops within and between varied social systems we inhabit affect action and thinking.

In Conversation Theory (Pask, 1976, p. xi), systems representing individuals are embedded in social contexts. Pask lists models of Vygotsky, Piaget, and Luria as approaches adopting such frameworks, using conversational methods treating action as the unit of analysis (the smallest part that retains the characteristics of developmental processes). The unit of analysis has been defined by Piaget in terms of the individual, using schemes that help conduct operations on the environment, and by Vygotsky in a dialogical setting, constituting observable, dynamic interactions that change mental models (Granott, 1998). Pask (1966, p. 226) infers one cannot specify to what extent developmental factors are affected by internal evolutionary processes or social interactions; this is a contention upon which Vygotsky and Piaget agree to disagree. Piaget suggests social experiences of humans recalibrate the canvas of stage-wise development; Vygotsky’s theory treats development as constructed from social processes. Discussing these nuances and magnifying focus on sociocultural theory, Pask (1976, pp. 19–20) describes how paired experiments, a method used by these scholars, involves an experimenter and participant working together on mechanical artifacts, manipulating them using contingent support from the expert party to understand the how and why of related concepts until the neophyte can provide an integrated explanation to the experimenter; what Pask (1976, p. 27) calls teachback.

Learners must clarify doubts and express views to negotiate learning (Scott, 2001). By engaging in collaborative exchange and building on each other’s knowledge, interacting systems act as novices and experts depending on situations, learning from one another. The idea of individuals (learners, instructors) and machines using networks of circular feedback loops to improve one another’s knowledge creates a mirror to Vygotsky’s zone of proximal development (ZPD) incorporating the capacities of responsive technologies. Pask’s teachback can be tied to obuchenie (an important aspect of the ZPD), which refers to learning from others and becoming “a head taller” by incorporating others’ experiences and knowledge into their own (Barrs, 2017). Imitating prototypical versions of tasks allows competence, but energetic emotions while performing tasks (e.g., writing a short story) may aid in developing skills and thinking. Vygotsky (1999) suggests the ZPD is the intellectual space lying at the crossroads of experiences/emotions and direct imitation. Those disseminating knowledge during learning take the student to a certain level of awareness, but to unearth new possibilities, the learner must recreate information through the lens of their own experiences (Glassman et al., 2022). Pask (1966) acknowledges that spontaneous experiences occur across truncated cultural systems; individuals are embedded within complex societies, as sociocultural psychology would suggest. The distributed internal monologue an observer engages in to construct meanings inspired much of second-order cybernetics, which suggests self-organizing systems constantly morph frames of reference by interacting with the environment (Scott, 2021). Laurillard (2002) has taken much influence from Paskian cybernetics in creating her conversational framework involving conceptual/demonstrative feedback loops of distributed peer and teacher communication, bringing conversants to new understanding and thinking within a ZPD.

Pask defined cybernetics as the “science or the art of manipulating defensible metaphors, showing how they may be constructed and what can be informed as a result of their existence” (Tzafestas, 2017, p. 124), and applied conversation theory to understand how systems (living systems and machines) respond to emergent perturbation in the environment through the terminology of M-individual and P-individual. M-individuals are brain body systems or processors/hardware for mental models/software. P-individuals are psychosocial (e.g., a perspective, belief, a social role). A single M-individual can embody several P-individuals or mental models. Several M-individuals may also embody a shared psychosocial activity (Scott, 2021, p. 28) and exteriorize mental models through language or L (see Figure 1). L is executed by the psychosocial P-individual in terms of concepts (L¹), but is given time and place in the physical realm, as a practical activity (L⁰; Pask, 1975, p. 72).

OPEN IN VIEWER

Figure 1.

M- and P-individuals.

Both P-individuals and M-individuals are coupled with social environments that fuel self-producing concepts in these organizationally closed systems. Structural coupling of social and biological developmental factors reflects Pask’s (1981) elaboration of biologist/cybernetician Humberto Maturana’s autopoiesis. Maturana suggests human nervous systems comprise closed physiological feedback loops but are also coupled with the sociocultural world (Mingers, 1991). Structural coupling of living systems with the environment aligns with Vygotsky’s assertion that phylogenetic/biological processes and maturational/ontogenetic factors are affected by cultural tools and symbols (which mediate learning) and specific/microgenetic events to different degrees (Marginson & Dang, 2017; Scribner, 1985). Pask draws the notion of multiple psychosocial unities or “gaggle of P-individuals” being embodied by groups of brain–body systems from Vygotsky (Pask, 1975, pp. 402–403), and asserts sociocultural phenomena become his focus.

All parties in conversations may enrich perspectives, using new knowledge to augment their realities and achieve goals. Nuances supplement mental representations as new situations unfold (Pask, 1975), taking conversations in open-ended directions and deepening critical thinking. A strict conversation (e.g., teacher–learner interaction) may involve less experienced individuals asking experts how and why questions, requesting/performing demonstrations and verbal evaluations, possibly guided by a machine. Natural conversations are potentially endless and may exist across psychosocial systems (families, schools, organizations) (Scott, 2021). Concept building is assisted by what Pask calls an entailment structure, or networked nodes representing concepts and subconcepts that may be studied (Pask, 1976, p. 51). When cyclical relationships between concepts are added to explain them in terms of one another, it is called an entailment mesh. While entailment structures represent the conceptual why of knowledge, a task structure represents the how or what may be practically done with these concepts (Scott, 2001). Ideas are unpacked or unzipped through varied epistemological lenses: holist (in terms of global concepts), serialist (in terms of ordered concepts), or versatile (a hybrid epistemology). These lenses may change based on the nature of sociohistorical experience, what Vygotsky (1999) terms perezhivanie. By representing concepts in dynamic webs, Pask treats them as processes that bring about, recall, or maintain relationships between topics (Pask, 1975, p. 553).

In Pask’s cybernetics, the machine becomes a mediating interface as in Vygotsky’s theories of development (Penuel & Wertsch, 1995), and additionally may serve as a responsive agent in conversation. Conversation theory may become an information age mirror to Vygotsky’s approach, capable of use to explain qualities of the responsive feedback loops of ubiquitous technologies like the internet in guiding human–human, human–machine, and machine–machine conversations. In the next sections, we delve into the roots of Pask and Vygotsky’s work, which show common ties to the theater scene in Britain and Russia, and interpret Pask’s aesthetic technologies through Vygotsky’s lens, tying the ZPD and perezhivanie to conversation theory.

Paskian aesthetic technologies through a Vygotskian lens

By the time Pask met Wiener and decided to create machines that augment theatrical performances, he was contributing to surrealistic light shows in Cambridge and London (Pickering, 2010). Pask created mechanized “aesthetically potent environments” for performers and audiences, incorporating technology into the British theater scene. In Pask’s aesthetics, observers abstract visual information and create internal conversations from works of art, which may change based on social experiences (subsequent conversations with another viewer or seeing another painting may change interpretations). Audiences and painters may undergo ideological change based on encountered events, reformulating artistic interpretations. Pask suggests his aesthetic environments fulfilled four affordances (Pickering, 2010):

1. They provide controllable variety or features that guide activity towards certain outcomes

This criterion is derived from Ashby’s law of requisite variety (Umpleby, 2009), which inspires models of self-regulation (Bandura, 1993). The law, explained by Ashby (1956) in Introduction to Cybernetics using the example of turn-taking games, suggests that to navigate the disturbances a social world produces, a system must have a set of responses at least as nuanced as these disturbances. Accordingly, the variety offered to users by technology must be at least as nuanced as their users to guide their thinking/interaction.

2. They contain information for humans to abstract and evaluate

This criterion deals with the intake of knowledge, aligning with aspects of information processing approaches.

3. They provide tacit cues defining activity or praxis based on affect and interest

This criterion suggests by spurring collaboration and discourse, cybernetic machines may offer context-specific responses expanding possibilities associated with activity and behavior, operationalizing a Vygotskian (1999) approach treating action as the unit of analysis.

4. They engage in ongoing collaboration with observers and adapt themselves

This criterion accounts for moment-to-moment responsiveness of Paskian machines to group/individual human agency.

The four criteria provide a multilevel approach housing social-cognitive, and sociohistorical approaches to mental models that consider both latent and observable characteristics of human functioning, and may be used to create qualitative coding schemes to decode characteristics of observed technology usership in naturalistic environments. The epistemology of Paskian cybernetics, and his machines, situates ideation/abstraction in dynamic agency and action, much like Vygotsky’s approach, suggesting that exteriorized behavior may change experiences based on environmental constraints and affect the mind simultaneously (Pangaro, 1993; Pask, 1976, pp. 352–356). Internal monologues and social action become one integrated whole, concomitant with the sociocultural approach.

Vygotsky portends human behavior and thought work similarly; ideas that came to him as he wrote theater reviews as a member of the “Khudjestvenny Sovet” (Art Council) in Russia, which granted access to plays in Gomel, and opportunities to be part of reading groups with directors like Vsevold Meyerhold. His early essays drew from Konstantin Stanislavski’s work, and understood relationships between external and internal planes of the actor, rooting them in social actions (Sobkin, 2016). Tom Scholte (2016), a contemporary actor and cybernetician who employs Pask’s theories, suggests Stanislavski’s naturalist system is inherently cybernetic, allowing performers to embrace their experiences in the present moment and adhere to a script. Scholte calls for reincorporation of spontaneous affect and experience leading to the emergence of “genuine” behaviors under imaginary circumstances, further amplified by peer feedback during rehearsals; actors become observers constantly reconstructing semiotic sequences of performance. He suggests such performances and rehearsals enable becoming an authentic vessel for a character, being oneself but using words one may not always utilize.

In line with Pask’s and Scholte’s writings, Vygotsky’s suggested affect was an important part of theatrical craft, highlighting how emotional activity with mediating tools affected actions and development of thinking (Smagorinsky, 2011). In later works like Mind, Consciousness, the Unconscious (1997), Vygotsky analogizes critique Stanislavski had for actors and directors (saying they focused on results rather than processes of acting), believing connections between mental functions could not be understood by separating cognition and action, relying on input–output driven rationality. Social action becomes the main indicator of the development of thinking in Vygotsky’s work. Each action could be separated into what Vygotsky calls “bits”; an irreducible part of the whole (he uses the example of water molecules, suggesting their properties can be understood looking at holistic molecular structures, not individual atoms; Vygotsky, 1997). Externalization of speech and internalization/reformulation of ideas are treated as united processes.

Many researchers whose ideas descended from Vygotsky (Engeström & Sannino, 2021; Kozulin, 1996) affirm that action is the primary unit of analysis of mental models, outlining how development is mediated by social experiences, past and present. Vygotsky refers to this idea as perezhivanie (derived from “pere,” through, and “zhivat,” to live; Vygotsky, 1994). Perezhivanie is linked to Stanislavski’s conception of acting methods, which focused on an emotional (re)experiencing of the lives of characters in historical contexts combined with the current temporal moment in the actor’s life (Glassman et al., 2022).

Vygotsky extends Stanislavski’s (2016) ideas beyond affect, outlining perezhivanie as ongoing interplay between personal/constitutional characteristics and situational/environmental phenomena. Experiences inform thoughts/beliefs/emotions, deciding exteriorized behaviors and subsequent reformulation of mental models (Michell, 2016). In the next sections, we discuss how Pask’s work incorporates the possibilities of human experience through a discussion of his aesthetic machines.

Musicolour and perezhivanie

One of the first devices he created, Musicolour, was inspired by Pask’s interest in synesthesia and helped gauge if a machine could adaptively learn relationships between sound and visual patterns to enhance a singer’s performance by co-operating with them (Bird & Di Paolo, 2008). Musicolour produced photoelectric output using relays (operationalized as a color wheel to select visual patterns) when sound input exceeded certain thresholds. If the performer sang monotone tunes, the device would stop responding, getting bored.

Musicolour allows performer and machine to learn from one another (see Figure 2; Pickering, 2010). The device produces disruptions in performance, prompting performers to think they need to add more thrust to their vocals. Performers form a shared language with the machine, to get accustomed to the mood of their performance, but establishing this connection requires the performer to play an interactive game with the machine. As they learn visual patterns the machine selects, performers adaptively produce certain notes to create light shows they know audiences have enjoyed (Pask, 1971).

OPEN IN VIEWER

Figure 2.

Musicolour.

When a performer uses Musicolour, initial actions, or notes sung, depend on experiences rehearsing the song, and the situation (the production being staged). The machine responds to input by producing visual effects. To keep the light show entertaining, the performer must advance their craft/performance and make the machine respond in a way that is pleasing to the audience. Both performer and machine exchange nuances about the how and the why of the performance through an ongoing conversation (Pangaro, 2008).

The interplay between performer and Musicolour represents shared consciousness between human and machine. This is akin to the conversations between Kitty and Levin (derived from Anna Karenina; Tolstoy, 1878/2014) used by Vygotsky (1934/1987) as an analogy for the inseparable link between cognition and affect in learning. Levin, who shows limited capacity for communication, relays feelings for Kitty using coded language incorporating the first consonants of words (she can understand him owing to their experiences together). Just as shared activity between Kitty and Levin has concrete meaning within their system of interaction (Glassman, 1996), the shared consciousness of Musicolour and performer allows the performer to gradually attune themselves to the nature of the machine’s communication and vice versa. The performer’s present emotional experience (of the performance), their cognition/thoughts (prompted by machine response), and their history (experiences rehearsing the musical piece) decide the nature of the resulting performance. Personal and situational factors guiding accumulation of experience mediate exteriorized action. Perezhivanie forms an integral part of the machine’s workings.

Cybernetic theater as a zone of proximal development

During the 1960s, Pask collaborated with theatre director Joan Littlewood (Werner, 2018), to think about ways to reformulate plotlines for plays based on audience responses, producing the proposal of a “Cybernetic Theater” (Pask, 1964). The proposed framework allowed audience members to signal to performers to enact certain sequences/scenes to change plotlines of a production to fit their preferences. Information would be relayed from audiences (who picked endings and plotlines from a range of choices) to interpreters, who signaled to actors on stage via headphones or earpieces to act out certain scenes (see Figure 3). Actors would rely on “metainformation” in their rehearsals and character development, and reformulate it based on audience responses, creating shared agency between performers and audiences (Pickering, 2010).

OPEN IN VIEWER

Figure 3.

Cybernetic theater.

While the play had a rehearsed structure, spectators could create circular feedback loops with actors through crowd-generated feedback (Pask, 1964). Nonhierarchical communication between audiences and actors culminate in shared narratives operationalized as a collaborative, evolving game. Actors would start influencing the audience’s thoughts through exteriorized behaviors, prompting spectator responses to build the play. Pask’s cybernetic theater transcended theatrical narratives that express pregiven plotlines, involving audience participation to encourage actors to experiment with unforeseeable outcomes (Pickering, 2010). Integrating distributed social roles undercut existing power structures in implementing theatrical performances.

The cybernetic theater’s functioning is tied to Stanislavski’s work, which aimed to achieve unity in acting, overcoming the predicaments presented to actors in Diderot’s paradox (Vygotsky, 1999). Diderot suggests actors performing solely from passion and emotion often lack road maps to replicate performances in good faith. Scientifically imitating a script, when applied in isolation, may lead to precision, but could limit character development. Stanislavski suggested actors could hone their craft at the crossroads of emotional experience and scientific precision. Vygotsky extended Stanislavski’s theatrical system, applying it to the development of thinking and affect during learning. The cybernetic theater is emblematic of a Vygotskian ZPD; an intellectual space lying at the fork between following mandated information and responding on the fly to the atmosphere of the present situation. Pask, who read some of Stanislavski’s work, may have drawn some of his ideas about performance art from this line of scholarship (Müller, 2016).

The cybernetic theater allows information to be recreated at the group level, not only dialogically. Rather than presenting knowledge top-down, the device allowed crowd-generated opinions to affect outcomes in theatrical performances through dynamic conversations. Crowd-generated knowledge-building is also supported on today’s collaborative internet platforms (e.g., social media) allowing human-to-human collaboration (Werner, 2018). However, rather than providing digitally indexed information to users like today’s internet forums allow collaboration, the cybernetic theater functioned as an analog device, providing contextual support to a theatrical performance. While the device was never built, its wiring diagram formed the inspiration for Pask’s other artistic endeavors, like the colloquy of mobiles, which understood human adaptation and evolution as occurring within highly interconnected social environments.

Colloquy of mobiles as a distributed system

In the 1960s, the Institute for Contemporary Arts became a hotspot for aesthetic creations. Assistant director Jasia Reichardt aimed to display intersections between technology and art, culminating in an exhibition called Cybernetic Serendipity. Pask created an installation, the Colloquy of Mobiles. The device comprised five robots, three “females” (designed by famous ballet designer and painter, Yolanda Sonnabend) and two “males” (designed by Pask) suspended from a ceiling, equipped with mirrors, lights, and sensors. They rotated on an axis, searching for “mates” to satisfy their drives (O and P drives, represented by orange and puce light; Pickering, 2010). Males would rotate on a 180-degree trajectory. Locating a female would lead the male and female units to perform a mating dance. The process would continue unless the machine was disturbed, or unless the male began to seek satisfaction of a different drive. Cybernetician Paul Pangaro created a functional replica of the Colloquy of Mobiles in 2018 in collaboration with students as part of a project-based endeavor in the Interaction Design program at the College of Creative Studies, Detroit (Pangaro & Mcleish, 2018).

Activity in the Colloquy of Mobiles represents adaptation or satisfaction of basic biological drives as a socially driven, open-ended process. This runs parallel to Vygotsky’s idea of natural (phylogenetic) and individual (ontogenetic) development as a process nested within the sociocultural world, subject to emergent events/information (Marginson & Dang, 2017). As outlined by Scribner (1985), phylogenetic development in the natural world leads to displacement of purely biological drives by characteristics of the sociohistorical context. Pask (1976) calls this the P-individuation of an M-individual.

In the Colloquy, ontogenetic actions of individual machine units are driven by both “desire to mate” and configurations/social situations offered by the distributed system of interacting units. In the phylogenetic development of the system and the satisfaction of all drives to completion, the tendency to perform a biological “mating dance” is displaced by the movements or social actions performed by moving units. Human observers interacting with the installation incorporate their experiences into the mating dance (see Figure 4) and change the courses of action taken by the machine to produce pleasing visual effects (Pickering, 2010).

OPEN IN VIEWER

Figure 4.

Colloquy of Mobiles.

Distributed feedback loops are formed between machine and spectators, driving the machine to search for new pathways to meet its desires, acting based on previous experiences in the system. The Colloquy echoes Pask’s (1968) assertion that humans seek novelty in their lives. They may take information from the social environment and use it to inform new behavior and thought. This falls in line with Bush’s (1945) web of trails, which outlines how humans may interact with technologies, gain information, share it with others, and make analogies to their own experiences to decide further actions and thoughts. Bush’s conception of distributed systems powered the internet’s creation and foreshadowed personal computing (Tilak et al., 2022). Present internet algorithms provide information streams based on corporate interests. These technologies may differ from devices providing access to the how and why of information stressed upon by Pask, becoming instruments of social control straying from the intent of second-order cybernetics. We suggest that deeper understandings of the capacities of technology across wider social systems may help to better understand how to decode the effects of postmodern technologies on society. Using the Fun Palace, an architectural endeavor Pask contributed to, as an example, we understand how interactive public spaces may form an apt metaphor for mosaics of computer programs that make up the fabric of the internet.

Fun palaces as virtual architecture

Apart from finding application in targeted spaces like theaters and art galleries, Pask’s work had architectural significance, finding application to public spaces. Joan Littlewood and architect Cedric Price collaborated on interactive installations housed within minimalistic spaces where individuals could gather on leisure days to learn and play together, to participate in aesthetic experiences that may spur personal development. Littlewood wanted to create a new kind of theater (Mathews, 2006) where people could experience new ideas and activities as players rather than as audiences.

The proposal of the cybernetic theater, through which Pask (1964) expressed interest in the idea of the Fun Palace, offered a simplified taste of the large-scale, dynamic installations Littlewood envisioned. Pask was invited to head the Fun Palace Cybernetics committee, to help understand how responsive technologies and interpretations of mental models could be used to iteratively redesign the structure. A scaffold-like armature inspired by Bauhaus architecture was envisioned, to house constantly changing art and activities, rather than prepackaged exhibits for a generalized public. Pask suggested incorporating information technologies to understand usership of the space. He wished to create environments that could meet the changing needs of a human population and allow engagement in a variety of activities (Mathews, 2006). The proposed design comprised sensors and response terminals to gather data from users, and an IBM 360-30 computer to quantify user trends. Based on this quantification, walls and staircases of the system would reformulate themselves to produce more varied experiences (see Figure 5).

OPEN IN VIEWER

Figure 5.

Fun Palace.

While Pask’s contributions to the Fun Palace committee were regarded as a form of social control through technology, he tried to incorporate the idea of iterative design, rather than a prescriptive packet of exhibits and activities into the Fun Palace. The mosaic of programs of which the structure was comprised enabled it to organize itself to respond to incoming human agency and meet needs of visitors to heighten aesthetic experience. Virtual architecture of the proposed Fun Palace could reprogram itself using escalators, travelators, warm-air screens, ultraviolet light, and optical barriers, and even “clean itself” through the circulation of river water. The construction of the Fun Palace never ensued. Interference from planning authorities in London led to its dissolution. Its initial floor plan was morphed and used as the prototype for the Centre Pompidou in Paris (an integrated space housing a museum, public library, and a center for music and acoustic research).

The Fun Palace provides a physical representation of how the modern technologies we use on the internet may be (re)designed based on understanding the needs and wants of consumers on an ongoing basis. However, algorithms of the present internet, rather than solely understanding how to augment the thinking of users, often incorporate prescriptive corporate interests that change the intent of such tools. In the next section, we outline how looking to second-order cybernetics may provide deeper understandings of the role(s) of ubiquitous online technologies and the processes involved in their design.

Paskian devices and the internet

Conversation theory focused on treating learning as a competitive and co-operative conversation between machines and those performing operations on machines within moving social environments (Pickering, 2010). Pask and Vygotsky both focus on balance between faithful imitation to acquire knowledge and acting on one’s emotions to perform a task in alignment with historical experiences. In the ZPD, or space between these facets of thinking/action, personal experiences and situational/social factors (perezhivanie) prismatically recreate activity (Michell, 2016). In Vygotsky’s theory, tools often become mediators, whereas Pask treats machines as surrogate participants in conversations with learners, supporting networked concept development. Second-order cybernetics, which considers living systems as observers constantly reconstructing their reference frames in navigating reality, and inserts researchers into the equation as participant observers (Scott, 2004), provides a robust framework for understanding collaborative conversational processes. The idea of cyclical collaboration formed the initial blueprint for nascent online communities like the Whole Earth Lectronic Link (WELL). The WELL’s Bulletin Board system enabled accessing communities or conferences of one’s interest and spurring critical discourse about topics ranging from parenting practices to the music of the Grateful Dead (Rheingold, 1994). Rheingold outlines how experiences in the parenting conference exposed him to support systems hosting discussions about issues like leukemia and how it affects a child’s daily life, using the interface. However, today’s internet portrays a different story. While it is chameleonic and offers a sandbox for knowledge uptake and collaboration, algorithms dominating the internet have been influenced by psychology’s focus on reverse engineering human thinking (Tilak & Glassman, 2020), pivoting commercial interests to control human activity.

While Paskian algorithms have been generalized to create hypertext-driven internet platforms and intelligent tutoring systems (ITS), there is scope to reincorporate them without the influence of cultural markets as a major part of the architecture of the internet by borrowing from Paskian cybernetics, to enhance contextual possibilities of technology-supported learning and collaboration. Understanding the functionality of present webtools using Pask’s framework may offer insight into how we may use a cybernetic framework to decode the effect of present online technologies on human activity, direct human thinking, and action to navigate present online feedback loops productively and redesign present tools by inserting researchers/designers as participant observers in crafting online tools who interpret changing perspectives of users on an ongoing basis.

Contemporary web technologies designed to take individuals towards a mandated skill level at games like chess parse knowledge to decide courses of action to oppose users. The online computer uses a decision-tree approach to make approximations of a set of the strongest moves for a fixed skill level algorithmically (Ensmenger, 2011) rather than based on the dynamic skill level of the player. The support provided to the user is continuous, but not rooted in the present reality (dependent on perezhivanie or accumulated sociohistorical experience) that the user inhabits in their learning. ITS, which lie at the crossroads of cognitive science, education, and technology, contain four models: (a) a domain model containing knowledge of a field, (b) a knowledge model of teacher strategies, (c) a model of learner’s decisions, and (d) a communication model to relay information.

The symbolic approaches used to construct ITS may provide contingent support and information but may not lend themselves to the adaptive capacities of both adaptive hypertext systems (AHS), or even analog technologies (Wilson & Scott, 2017). Pask’s Solartron Adaptive Keyboard Instructor (SAKI) provides a sound example of analog technologies used for teaching simple skills. It enabled learners to bridge knowledge gaps in typing skills by providing exercises of varying difficulties based on the current competencies of the user (see Figure 6).

OPEN IN VIEWER

Figure 6.

SAKI.

SAKI would identify faults and the current skill level of the participant by understanding the quality and response time of inputting data into a Hollerith Key Punch pad. It provided a set of typing exercises that matched these parameters, dimming its display as exercises were made more challenging (Romiszowski, 2013). SAKI’s wiring was applied to create more advanced machines using digital models to simulate and guide learning, like EUCRATES (which allowed teacher and pupil simulations to understand how teaching and learning lead to several open-ended possibilities), used to model typing tests and even fighter pilot simulations. Teachers could model learning scenarios by changing the external and internal awareness of the machine, its obstinacy and willful avoidance of learning, and even increasing focus on memory reinforcement. Results from simulations could be used to model teaching like a game of strategy. SAKI and EUCRATES allowed individuals to organize activity by trying to meet standards set out by the machine, but always provided room to understand where users were in their teaching or learning to provide opportunities for gradual development.

Distributed interplay between machine and human systems on the internet may encourage wider possibilities for open-ended conversations. Wishing to investigate human–machine conversations that formed the blueprint of the internet, Pask began creating technologies that could respond autonomously to users using algorithms like ITS, and link existing ideas to new ones using hypertext (Wilson & Scott, 2017). These hybrid devices, like CASTE (Course Assembly System and Tutoring Environment) enabled creating ongoing understanding of topics based on learning strategies/epistemologies of individual learners and conceptual/demonstrative responses to machine-provided information. CASTE had four key features: (a) an entailment structure displaying a web of concepts related to a particular subject, (b) a modelling facility to demonstrate and aid understanding, (c) a Belief and Opinion Sampling System (BOSS) to sample mistakes and uncertainties about topics, and (d) a communication console to be run by the student allowing them to state an aim, choose a topic, ask for explanations/demonstrations, or give them (see Figure 7).

OPEN IN VIEWER

Figure 7.

CASTE.

The learners could explore a topic if they displayed some prerequisite understanding of it, and then could elicit or provide goal-directed demonstrations and explanations to support independent performance. They would pick topics from an entailment structure to grasp a deeper understanding of concepts, with the system acting as a surrogate participant in the learning conversation. At the end of the class or learning module, learners would be asked to display their overall understanding of chosen topics; a satisfactory explanation would be indicated by a green light illuminated above a concept in the entailment structure, controlled by the experimenter or tutor (Pask, 1975). While responses of CASTE were algorithmic, they expanded the activity of a learner by providing tutorial strategies to expand understandings in open-ended ways.

Furthering the framework used to create CASTE, Pask wished to craft a more flexible and powerful system (Pask & Kopstein, 1977) that could enable both individuals and groups of learners to create an epistemological laboratory, taking each user on personalized learning journeys (Pangaro, 2001) to build entailment meshes and to specialize them as entailment structures through the assistance of experts/authors of tutorial materials. THOUGHTSTICKER offered tutorial messages that contained links to a neighborhood of topics in a class’s entailment structure (e.g., clicking “biology” may unzip further topics about cells or evolution). These links offered pathways to other concepts covered in a course, enabling learners to further augment understandings of topics in an expansive fashion. Rather than underlining hyperlinks as per usual, the system allowed hovering over words to show entailments derivable from head concepts. Back and forward buttons allowed browser-like navigation, and the string of topics was determined by a student’s learning profile (serialist, holist, versatile) and usership history. A jumping function was also provided to allow choices to move to more distal related concepts. Students’ contributions to an entailment (agreements, disagreements, novel findings) could be tagged back to them, creating circular links between students and the device (see Figure 8).

OPEN IN VIEWER

Figure 8.

Thoughtsticker.

By providing cued responses to students and encouraging open-ended conversations by using recommendation mechanisms deriving insights from neighborhoods of conceptual links discussed by learners, CASTE and THOUGHTSTICKER can serve as cybernetic ZPDs. Other technologies like Byzantium, which has been used to teach/learn accounting, provides step-by-step guidance but may not link neighborhoods of concepts to create personal computing experiences for learners, or be applied in a domain-general manner (Patel et al., 2001). Popularized knowledge-sharing tools like Knowledge Forum (Scardamalia, 2004) have rekindled the ethos of the web of trails in the educational sphere. Scardamalia (2004) provides an example of elementary students studying paleontology who created classifications of their favorite dinosaur species using graphics. These were supplemented by college students who added allusions to geological time into the mix of ideas. While Knowledge Forum may spur collaboration, it does not provide autonomous responses like CASTE and THOUGHTSTICKER did, and their functionalities stand apart from machine feedback loops of internet tools. While robot tools supporting collaboration by regulating the variety of interactions have also been created (Nasir et al., 2022), they provide finite feedback that may not simulate the ubiquitous influence that personal devices have on our activity and thinking.

Algorithms making up digital Paskian technologies like CASTE and THOUGHTSTICKER have, today, been generalized to create the modicum of webtools that have become a ubiquitous part of society (Wilson & Scott, 2017). Present social media platforms like Facebook, Twitter, and Instagram may offer opportunities for synchronous communication/collaboration and use recommendation systems to create explore feeds and timelines for users based on their frequent interactions by gauging the quality or domain of the content users are interested in (Pangaro, 2008). However, commercial algorithms may subvert social influence increasing the uncertainty of human functioning in order to heighten accuracy of observations of human thinking (Chavalarias, 2016). Suggestions provided by these systems are based on keyword frequencies rather than context-specific conceptual links created by a group of users. Pask’s (1980) distinction between communication (making links with others) and conversation (exchanging epistemologies and ideologies) helps understand why swathes of online information create an “illusion of togetherness” that hampers critical dialogue.

Looking to the framework of Paskian cybernetics may help understand mechanisms to reorient technology usership towards productive collaboration. Conversation theory suggests that human–human, human–machine, and machine–machine interaction may occur across varied social systems (Scott, 2021). Understanding how we use informal online technologies on personal devices through user reports of thinking and action resulting from online feedback loops may help researchers understand how to direct usership of such technologies towards learning and online community formation (Tilak & Glassman, 2020).

The second purpose of a cybernetic framework in the Information Age would be to provide cues to redesign current online technologies. Recently, Pangaro (2008) applied Paskian algorithms to creating software spurring reflection during online reading. Thoughtshuffler points out keywords within articles it feels users may be interested in, giving them a personalized gist every time they decide to engage in scrolling. The more qualitative feedback loop produced by Thoughtshuffler aids a reader depending on their own individual ideologies and preferences based on a neighborhood of related concepts a user is interested in, rather than redundant information based on frequently accessed keywords. However, with the internet moving towards quick searches for information, processes of finding information that may develop one’s thinking in open-ended ways are rarely considered by those designing tools that affect human thinking and activity (Tilak & Glassman, 2020). Commercialized enterprises selling goods, producing web copy, and designing search engines focus on understanding trends in consumer behavior and guiding them towards prescribed packets of contingencies. The intent of the designer/observer plays an important role in power structures of cultural production and media (Habermas, 2006). Pask’s conversation theory, which provides a framework for interaction design (Pangaro, 2008), suggests that the breakdown of such power structures begins with creators iteratively changing tools based on gathering qualitative experiences of users in real time through researcher–user partnerships.

With the advent of conversational assistants (e.g., Alexa, Siri) designed by large corporations, it becomes easy to ask for what one wants using portable devices. Pangaro (2016) uses the “parable of Luigi’s Pizza” to explain this phenomenon, saying Alexa would tell someone what pizza a restaurant has by approximating whether one likes it, whether it is economically feasible ordering a slice. It would not explain how/why it concluded the user would enjoy it or provide personalized results based on real experiences. This logic is opaque to users, with companies designing algorithms fighting to safeguard code, and aiming to guide thinking towards linear results through instant gratification. Providing information to users without opportunities to critically analyze it and expand ideas through deeper reasoning may lead to concerning outcomes (e.g., conspiracy theories, Walther & McCoy, 2021; and maladaptive mental health outcomes, Haidt & Allen, 2020) often tied to online activity.

Participatory design and using Paskian algorithms without the veneer of corporate interests to expand the capacities of today’s online tools may tap into possibilities of using technology to expand human thinking and interaction. Such a design framework may be achieved through a qualitative understanding of dynamic user experiences on present technologies to provide cues to redesign tools. Borrowing from conversation theory and the techniques used to design Paskian technologies may enable individuals to learn to use present online tools to take activity towards a critical eye for information, and help designers reorient intentions behind these tools. A cybernetic framework applied to psychology and the digital humanities may equip society with new opportunities to navigate the information age.

Conclusion

With online tools creating a networked public sphere, there are hopes and dangers with internet use (Tilak & Glassman, 2020). Digital humanists like Marcuse (1941) warned of technological rationality, which streamlines human thinking to meet mandated standards or outcomes (e.g., media controlling information and misinformation). A power imbalance in creation of tools may lead to the stifling of open-ended possibilities of critical reflection to navigate social realities. While providing sandboxes for knowledge creation (COVID-19 communities on Reddit, for example; Lai et al., 2020), online technologies may simultaneously create misinformation and concerns in larger social systems, forming a feedback loop for which we are not necessarily prepared.

We suggest Pask’s conversation theory may help psychologists, designers/technologists, and digital humanists understand context-specific roles of the internet in human development by helping to gauge how recursive feedback loops of technologies play surrogate roles in changing human thinking as we scroll through our personal devices moment-to-moment. Apart from providing an approach to observe and study (formal and informal) technology-assisted discourse and concept development across social systems, Pask’s framework also has implications for design. He suggests ongoing interplay between designer/observer and user to align role(s) and purpose(s) of tools with the needs of current contexts (Pangaro, 2008). Analyzing consumer activity responding to context-specific feedback loops of technology to inform processes like learning and community-creation may help humanize intentions of large-scale corporations, undercutting power structures arising from technological rationality by equipping individuals with strategies to navigate endless seas of information. Using qualitative methods (interviews, coding schemes derived from Pask’s technology affordances) to continuously understand usership may supplement the iterative design of online tools.

Further reinforcing links between Vygotsky’s cultural historical psychology and Pask’s cybernetics reveals how algorithms designed for contextual support may advance the capacities of the internet to meet visions and aspirations behind its creation.