Introduction

---

Throughout my life, ideas around music, computation, and cognition have been important to me, acting as almost invisible threads linking the various disciplines I have been passionate about since a young age. As part of a generation that grew up with digital computers at the reach of a hand, my artistic practice has been deeply shaped by digital technology.

After many years dedicated to piano performance, experimental music composition, and live electronic improvisation, I recently felt drawn back to academia, driven by a desire to better understand the “special” brain of musicians and what occurs during the act of music-making.*Or at least, that was the plan. In the end, my master's research project looked into aspects of musical perceptual expertise, in particular in tasks related to speech perception: Juan S. Vassallo, "Music expert-novice differences in speech perception," University of Victoria Repository Dspace (2019) https://dspace.library.uvic.ca/handle/1828/11407 As I delved deeper into this new field of knowledge, I began noticing intriguing conceptual parallels between the workings of the human mind and computational concepts that have been familiar to me since a young age. For instance, I observed that research trends in cognitive psychology often aim to formalize cognitive processes as systems, modeling the mind computationally, which often involves using mathematical or statistical models, such as Gaussian distributions, Bayesian inference, fuzzy logic, etc. I found these intersections fascinating.

“The central hypothesis of cognitive science is that thinking can best be understood in terms of representational structures in the mind and computational procedures that operate on those structures.”

Following these ideas, I naturally began to wonder whether computational conceptual models could contribute to our understanding of the creative process, especially in art, and in music composition, my primary field of practice. Can creative thinking be effectively captured through computational insight? Can even the most intuition-driven aspects of creativity be systematically analyzed and modeled? If so, could these ultimately be replicated by a computer?

I soon realized that these questions challenged the traditional –and somewhat romanticized– view of artistic creation as a mysterious, elusive process. What might the implications be for music creators? Could composers enhance their work by understanding the cognitive functions at play during the creative act? Or is it more beneficial for an artist to allow creativity and intuition to remain, as they are often perceived: inexplicable?

I felt that seriously pursuing these questions was worthwhile, even though other practitioners have explored these inquiries for some time. In my case, being a composer with a background in programming, and, more recently, cognition and brain sciences positioned me well to integrate these fields into a multidisciplinary project. Eventually, I formalized these inquiries into a doctoral project in artistic research with an interdisciplinary focus. This project would integrate theoretical discussions from these various fields with the development of my creative practice.

Outward Threads grew out of these early speculations, evolving into a collection of musical works accompanied by analytical, theoretical, and reflective writings that explore broader studies on human creative cognition, computation, artificial intelligence, and intersections between them. I have pursued this integration because I believe it offers valuable insights for practitioners who, like me, are interested in exploring the nature and mechanisms of the creative process, its potential connections to computational paradigms, and its manifestation in the act of composing. Importantly, this exploration does not overlook the social and cultural dimensions of a compositional practice, incorporating reflections on the ethical, institutional, and historical contexts that shape it.

This text and the accompanying artistic work are the culmination of my research at the Grieg Academy, University of Bergen, over the past four years. What follows is an exploration of the artistic outcomes and reflective insights of the project Outward Threads.

Outward Threads

“Daedalus’ maze is shrouded in secrecy. Can we open the labyrinth and count what is inside?”

Daedalus, a skilled craftsman and inventor in Greek mythology, was renowned for his architectural and mechanical mastery.*Among his many inventions, he designed the Labyrinth for King Minos of Crete to imprison the Minotaur. The title Outward Threads reflects a desire to contemplate the narrative of Daedalus as not only central to our conception of technique*Daedalus is proposed as an eponym for technique –or technical means– by Bruno Latour in "On Technical Mediation - Philosophy, Sociology, Genealogy." but also to that of artistic creation. In this light, the act of creation is seen as a process of exploration of a labyrinthic space within which the creative process unfolds, through the search and discovery across a structure of complex passageways. At the core of the labyrinth metaphor, therefore, lies the idea of exploration.

But what exactly is being explored? One could argue that the external world and our internal representation of it become the input for our creative processes. Theorists from various fields have conceptualized the idea of mental representations of concepts as a type of mental geographic space. A direct example of this is a theory of cognition by P. Gärdenfors, known as conceptual spaces.* I will discuss Gärdenfors’ theory of conceptual spaces in the section ‘Multidisciplinary Insights.’ In this view, exploring concepts in the mind can be metaphorically equated to navigating a mental space –whether best described as a labyrinth, a topography, or an ocean may be a matter of personal perspective. The notion of exploration as a creative mechanism is central to this reflection. The metaphor of the labyrinth ultimately crystallizes as a representation of the creative process for each work. However, it is not necessary to confine the idea strictly to the labyrinth; its use remains flexible.

In the opening of her book The Creative Mind: Myths and Mechanisms, Margaret Boden poses a quite thought-provoking statement:

“Computers and creativity make interesting partners with respect to two different projects. One, which interests me the most, is understanding human creativity. The other is trying to produce machine creativity –or rather machine “creativity”– in which the computer at least appears to be creative to some degree.”

As has become apparent, one of the primary interests of this research overlaps with Boden’s first project. This discussion is grounded in theories of cognition and will eventually explore connections with the field of Artificial Intelligence (AI). However, tracing these connections does not imply equating the creative process to the functioning of a computer. Rather, as Boden suggests, such comparisons can provide valuable insights into creative processes through certain analogies with conceptual frameworks from computation and AI.

The second project, however, concerning the creativity of computers or AI, cannot be entirely avoided. It surfaces occasionally throughout these reflections and will be addressed timely in this text. As the reader may be aware, AI –especially generative AI (gen-AI)– has emerged recently as an ubiquitous technological advancement. The astonishing results it produces in generating text and multimedia content are still a mystery to many. Potentially, and aside from too technical discussions, drawing analogies between these models and the functioning of the human mind could offer valuable insights to a broad audience on how these systems operate, and potentially, how they might evolve in the near future.

But coming back again to the idea of exploring a labyrinth, it can be argued that there are different ways of doing it. Some will do it following certain logical steps, others just following their intuition –or gut feeling. Maybe it is even possible to shift between these two: Some movements might require some logical consideration, others just feel necessary.

There are different ways of exploring a labyrinth. Somehow, I know this for sure, but I am tempted to ask ChatGPT if there are any 'logical' ways of doing it. This is a response from ChatGPT-4o to the prompt: “Is there any logical strategy to find the way out of a labyrinth?”

Yes, there are several logical strategies that can help you find your way out of a labyrinth. Some of these strategies are based on mathematical principles, while others rely on simple rules of thumb. Here are a few well-known methods:

1. The Right-Hand (or Left-Hand) Rule:
   How it works: Keep one hand on the wall (either the right or left) and continue walking, keeping your hand on the wall at all times.
   Why it works: This method ensures that you never revisit the same area, and eventually, you’ll find an exit. It works for any simply-connected maze (a maze without loops or disconnected parts). However, it might not work in a maze with loops or multiple disjoint walls.
   
   • Pros: Simple and easy to follow.
   • Cons: In mazes with multiple loops or islands, you may end up walking in circles around a central area without finding the exit.
2. Bread Crumbs or Marking Your Path:
   How it works: Leave markers (physical objects or chalk marks) at intersections or along paths to help track where you've been.
   Why it works: By marking paths you've already explored, you can avoid revisiting dead ends and make sure you try new paths.
   
   • Pros: Helps in any kind of maze, including ones with loops or complex structures.
   • Cons: You need something to mark the path with. In a very large or complex maze, it can become overwhelming to manage.
3. Pledge Algorithm:
   How it works: This method is useful when the right-hand or left-hand rule fails (e.g., when the labyrinth has loops or islands). The idea is to set a fixed direction (e.g., north), and after hitting a wall, you turn and count how many left or right turns you make, always trying to balance out your turns.
   • Pros: Works for more complex mazes with loops.
   • Cons: More mentally demanding, requires constant counting and awareness of directions.
4. Trémaux’s Algorithm (Backtracking):
   How it works: This is a more systematic approach where you keep track of intersections. When you reach an intersection, you choose an unexplored path. If you reach a dead-end or revisit a known intersection, you backtrack to the previous intersection and try a different unexplored path.
   • Pros: Guaranteed to find the exit eventually, and works for any type of labyrinth, including complex ones with loops.
   • Cons: Requires good memory or marking of intersections, and can take a long time to explore all paths.
5. Look for Patterns:
   How it works: Some mazes are designed with patterns that repeat or have symmetry. By observing these, you can often figure out shortcuts or predict where certain paths lead.
   • Pros: Can be very quick if the maze has a noticeable structure.
   • Cons: Requires keen observation and may not work in irregular mazes.
6. Maze Solving Algorithms for Computer-based Mazes:
   If you’re solving a maze computationally, there are several algorithms designed for that purpose:
   • Depth-First Search (DFS): Explores as far down a path as possible before backtracking.
   • Breadth-First Search (BFS): Explores all paths one step at a time, ensuring the shortest path is found.
   • Dijkstra’s Algorithm: Finds the shortest path in mazes with weighted edges.
   • A’ Search Algorithm: Optimizes search by using heuristics to focus on paths most likely to lead to the goal.

Summary

1. For simple mazes, the right-hand rule is quick and easy.
2. For complex mazes with loops, Trémaux’s algorithm or marking your path is a better choice.
3. If solving a maze computationally, use algorithms like DFS, BFS, or A’ search.ng A’ search.

Would you like to dive into any of these strategies in more detail?* “Is there any logical strategy to find the way out of a labyrinth?” prompt. ChatGPT 4, October 2023 version, OpenAI, retrieved Oct 15, 2024 from chat.openai.com/chat.

What if we applied the idea of logical exploration to a creative act? Consider a musical example: many creative processes involve following a sequence of logical steps or adhering to specific rules. For instance, composing a chorale or a fugue typically requires knowledge of tonal harmony and counterpoint, along with following a set of strict rules. By doing so, one is almost certain to produce a result. Perhaps even one of artistic value.

Of course, that’s not the only way music-making unfolds. Many musicians who improvise brilliantly or reharmonize songs in completely unexpected ways often say, “I have no idea what I’m doing.” These examples illustrate two seemingly contrasting approaches to musical exploration: one grounded in rational thought and methodical processes, and the other rooted in what we typically understand as intuition. Discussing these two modes of thinking –the rational and the intuitive– will play a significant role in this reflection. It is important to note, at the outset, that they represent distinct forms of cognition, a concept that will be explored in greater detail in the next chapter.

Margaret Boden further expands the idea of exploration in human creativity as the result of exploring, recombining, and transforming mental spaces.* I will discuss Margaret A. Boden's theory of Creativity on the chapter 'Multidisciplinary Insights' In her words, the act of merely exploring a space seems a somewhat straightforward task. Experimenting, observing the possibilities, experimenting again, going back to the starting point if it was not satisfactory, or we want to explore a different way, and so on. But the idea of recombining and transforming the space entails somehow that we can change it.

If we think again of the labyrinth, what would this entail? Perhaps jumping over the walls? Or breaking through them? The metaphor of the labyrinth may seem rigid now. However, I will maintain it, as it ultimately serves the purpose of illustrating the concept of exploration.

For Boden, the process of creative exploration mainly relies on mental strategies known as heuristics,* I will talk more about heuristics in the section 'The Conceptual Level', in the next chapter, 'Multidisciplinary Insights' which guide us in navigating complex mental spaces through a form of search. In the field of AI, the concept of search spaces has long existed as a method for exploring potential solutions to problems. This search space is sometimes traversed by employing algorithms or systems of rules; other times, it is navigated in ways that may not make much sense to the human mind. In the following sections, I will focus on these various exploration methods, using as a starting point some of the conceptual connections between human creative cognition and AI. I will examine how these explorations unfold for each piece, utilizing various methodologies that may or may not involve computational creative tools.

So far, it seems that in this text, I have put a large effort into describing mental processes. However, another interesting problem that I will discuss, in addition to mental complexities and mental representation, deals with the notion of embodiment and whether the body might have agency in the process of creatively exploring for creating music. Is the body a feasible tool for exploring creative spaces in music composition? How does a computational model fit the concept of embodiment, grounded on knowledge that is acquired and applied through action rather than through description or formal reasoning? The concept of embodied cognition is a relatively recent one which has found a stronghold in the study of cognition. However, it also raises questions and problems, which I hope to put into focus.

But isn’t this supposed to be an artistic research project? How does all this theory fit in, and where does the artistic practice come in?

Outward Threads is fundamentally an artistic research project focused on producing primarily artistic outcomes. It is important to emphasize that this project does not aim to conduct scientific studies or employ methodologies from non-artistic research disciplines. Accordingly, both its methods and results should be understood within an artistic research framework. However, as mentioned earlier, the project embraces an interdisciplinary approach, drawing on insights from various fields that inform the artistic work and expanding its inquiries beyond a strictly musical scope. The implications of these concepts are complex and potentially far-reaching. Therefore, this endeavor would benefit from a multidisciplinary perspective, presenting concepts from these fields in an accessible way that simplifies technical aspects for a diverse audience. While these reflections aim to delve into these theories, this text is not exclusively aimed at specialists in any of these fields. Furthermore, an important goal of this text is to make these concepts accessible to a wide range of readers, particularly those without a background in cognition, computer sciences, or even music.

---