Данная статья представляет собой научно-философский анализ детерминизма в контексте алгоритмических систем. В работе исследуется двойственная роль алгоритмов: как операциональных инструментов, организующих деятельность через функциональную предзаданность, и как структур, задающих познаваемые границы реальности, по определенным, заданным самими алгоритмами траекториям.
Ключевые слова: детерминизм, алгоритм, формальные системы, машина Тьюринга, причинно-следственные связи, непрозрачность, конструирование реальности.
The concept of algorithms as strictly deterministic systems that consistently and predictably transform input data into output data is fundamental for their practical application and theoretical understanding. The very concept of «algorithm» inherently implies strict adherence to a predefined sequence of actions. Here is an example of one of the definitions of an algorithm: «An algorithm is a recipe that is divided into discrete, elementary steps, providing a clear path from the variable conditions of a problem to the desired outcome» [1, p. 40]. However, upon deeper scientific and philosophical analysis, the determinism of algorithmic systems becomes more complex and problematic, addressing issues of knowledge, freedom, and the very nature of reality.
A classical algorithm based on formal systems, such as the Turing machine, is an «abstract model of an idealized computing device that combines formal and causal characteristics» [2, p. 62]. In this model, each step is a well-defined rule applied to the current state of the system.
The input data and the sequence of operations uniquely determine the result. This ensures reproducibility and predictability, which is crucial for scientific calculations, automation, and solving engineering problems.
The determinism of algorithms is known to be «the causal predetermination of actions and the unambiguousness of the result of executing a command given the initial data» [2, p. 62]. It reflects the reductionist tendency of scientific thinking, which is aimed at the explanation of complex structures by analyzing their simplest components and identifying universal causal patterns to the simple and predictable forms. Such predictability implies an objective reality that obeys laws, which is easy to understand, and utilize.
In modern algorithms, the problem of determinism is not clear. In his critical analyses of correlations, K. Meyasu concludes that reality is independent of human perception. This statement contradicts the idea of algorithms as deterministic systems. «From everything that can be mathematically understood in an object (in the form of a formula or in digital format), rather than from what is perceived or felt, it makes sense to make the properties of a thing not only as it is for me, but also as it is without me» [3, p. 9].
Based on this, the possibility of describing the world using algorithms in terms of determinism and predictability becomes problematic. The mathematical structures underlying algorithms simultaneously act as tools for knowledge (thus, we can talk about their subjectivity) and as properties of things (objectivity).
We can assume that algorithms can be a bridge between chaotic reality and attempts to organize it. An example of such attempts is ChatGPT. Based on a huge amount of data obtained from the Internet, the language model «answers questions, writes stories, scientific and journalistic articles, poetry, and even program code. ˂…> its amazing abilities and equally amazing mistakes are actively discussed on the Internet» [4, p. 10]. If we accept the postulate of the existence of a world «in itself», devoid of pre-established, necessary cause-effect relationships, then any algorithmic model acquires the status of temporary stabilization. It is not a reflection of absolute truth, but an ephemeral arrangement of a fundamentally changing reality [5].
Modern algorithms, especially those based on machine -learning, operate in an environment of exponential complexity. While each individual step may be deterministic, the nonlinear interaction of multiple parameters makes the result unpredictable and inexplicable, even to the creators of the algorithms (the «black box» problem), making its internal processes incomprehensible. This challenges the very concept of «knowledge» about the algorithm.
Algorithms, especially deep neural networks, are trained on data whose very nature contains fundamental uncertainty. The result of their work can be emergent. The result is hard to explain or predict based only on the properties of the system's constituent elements and their behavior. Often, randomness is deliberately introduced into calculations as a necessary condition for their operation, such as in genetic algorithms.
An algorithm ceases to be just a tool created to perform a specific task and becomes a subject that develops its own patterns, which may be unexpected or even contradictory to the creators' intentions. If an algorithm is trained on data that reflects social prejudices, its deterministic output will reproduce this bias under the guise of objective truth. This creates an illusion of impartiality, concealing systemic discrimination. The opacity and unpredictability of complex algorithms means that we may lose control over the systems that control our lives. Determinism is preserved at the code level, but human responsibility is lost. If we cannot trace the cause-and-effect relationship, we lose the ability to understand why things happened the way they did, to confidently predict the future behavior of the system, and to challenge its decisions.
Thus, we can conclude that the determinism of an algorithm is not always a direct reflection of objective cause-and-effect relationships, but rather a tool for constructing reality based on formal rules. «Algorithms materialize a philosophical paradox: as instrumental systems, they simultaneously constitute a new epistemological reality where order emerges from chaos. Their work confirms Meiyasu's thesis about a world without necessary cause-and-effect relationships, showing how modern science operates with chance, transforming it from a philosophical category into a working method of cognition» [5, p. 145]. If we start relying on algorithms to model and control the world, the question arises: where is the line between a model and reality? Algorithmic determinism, instead of reflecting reality, can begin to shape it, guiding it along specific paths determined by the algorithms themselves, with profound philosophical, social, and ethical implications.
References:
- Yurkov N. K., Rusyaeva E. Yu., Poltavsky A. V. A view of the theory of algorithms from the standpoint of philosophy. — Text: immediate // Reliability and quality of complex systems. — 2014. — No 2. — P. 40–45.
- Baryshnikov P. N. Metaphorical foundations of computerism in cognitive sciences and philosophy of consciousness. — Text: immediate // Philosophy of Science and Technology. — 2018. — № 2. — P. 61–72.
- Meillassoux, K. After Finitude: An Essay on the Need for Contingency. — 1st ed. — Yekaterinburg: Cabinet Scientist, 2015. — 196 p. — ISBN 978–5-7525–3069–2. — Text: immediate.
- Rezaev A. V., Tregubova N. D. From sociology of algorithms to social analytics of artificial sociality: analysis of API and ChatGPT cases. — Text: immediate // Monitoring of public opinion: economic and social changes. — 2023. — № 3. — P. 3–22.
- Kukel V. E. The reality of the algorithm — intentionality and instrumentality. — Text: immediate // Intellect. Innovations. Investments. — 2025. — № 3. — P. 141–148.
