Computationalism, Mind, and Artificial Intelligence

This article explores computationalism, also known as the computational theory of mind, as pioneered by philosophers Hilary Putnam and Jerry Fodor, and examines its foundational relationship to artificial intelligence. By defining the human brain as an information-processing system, this philosophical framework argues that human thought is a form of computation. This article breaks down the core concepts of the theory, the specific contributions of Putnam and Fodor, and how these ideas directly enable and influence the development of modern artificial intelligence.

The Core Concept of Computationalism

Computationalism is the philosophical view that the human mind is an information-processing system and that cognition and consciousness together are a form of computation. Rather than viewing the mind as a mysterious, non-physical entity or merely reducing it to raw biological brain tissue, computationalism sits in the middle.

The theory relies heavily on a computer metaphor: the physical brain serves as the hardware, while the mind functions as the software. According to this view, mental states—such as beliefs, desires, and thoughts—are computational states. Thinking is simply the process of manipulating mental symbols according to specific, logical rules, much like a computer processes data using algorithms.

Hilary Putnam and Multiple Realizability

Hilary Putnam introduced the modern computational theory of mind in the 1960s. He proposed that the human mind functions like a Turing machine—an abstract mathematical model of computation that manipulates symbols on a strip of tape according to a table of rules.

Putnam’s most critical contribution to this theory, and to artificial intelligence, is the concept of "multiple realizability." Before Putnam, many believed that mental states were identical to specific biological brain states. Putnam argued that a single mental state (like feeling pain or solving a math problem) could be realized by many different physical systems. Because the mind is defined by its functional organization (the software) rather than its physical makeup (the hardware), a silicon-based computer could theoretically possess the exact same mental states as a carbon-based human brain.

Jerry Fodor and the Language of Thought

Jerry Fodor expanded upon Putnam’s work in the 1970s and 1980s, becoming the most prominent defender of computationalism. Fodor sought to explain how the brain computes. To do this, he introduced the Language of Thought Hypothesis (LOTH), often referred to as "Mentalese."

Fodor argued that thinking occurs in a mental language that is innate and distinct from spoken languages like English or Mandarin. According to Fodor:

• Mental Representations: Thoughts are represented in the mind as symbol structures, similar to words in a sentence.
• Syntax and Semantics: Just like a computer program, the mind processes these symbols based on their syntax (their physical shape and structure) rather than their semantics (their meaning).
• Compositionality: Simple mental symbols can be combined in infinite ways to create complex thoughts, following strict algorithmic rules.

By establishing that human reasoning is the rule-based manipulation of internal symbols, Fodor provided a concrete blueprint for how a machine might replicate human thought.

The Connection to Artificial Intelligence

Computationalism is the philosophical bedrock of artificial intelligence. Without the assumption that intelligence is a computable process independent of human biology, the pursuit of true AI would be theoretically impossible.

The ideas of Putnam and Fodor directly birthed "Good Old-Fashioned Artificial Intelligence" (GOFAI), also known as Symbolic AI. For decades, AI researchers attempted to build intelligent machines by programming them with explicit rules and symbol-manipulation capabilities, directly mirroring Fodor’s Language of Thought. When an AI plays chess by evaluating board states (symbols) and applying movement strategies (rules), it is operating on computationalist principles.

Furthermore, computationalism distinguishes between "Weak AI" and "Strong AI":

• Weak AI suggests that computers can only simulate human thought.
• Strong AI asserts that a properly programmed computer does not merely simulate a mind; it literally is a mind.

Because computationalism states that thinking is computing, it inherently supports the possibility of Strong AI. If an artificial system can run the exact same computational algorithms as the human brain, computationalism dictates that the machine is genuinely thinking.

Criticisms and Evolution

Despite its massive influence, the Putnam-Fodor model faces significant challenges. Philosopher John Searle famously attacked the theory with his "Chinese Room" argument, asserting that a computer manipulating symbols based on syntax does not actually understand the meaning (semantics) of those symbols. Therefore, computation alone cannot produce true understanding or consciousness.

Ironically, Hilary Putnam himself later became one of the biggest critics of his own theory. He eventually argued that computationalism failed to account for how the mind interacts with the external environment, noting that meaning cannot be generated purely by internal symbol manipulation.

In modern AI, strict Fodorian symbolic AI has largely taken a backseat to connectionism—the use of artificial neural networks and deep learning. Instead of relying on pre-programmed rules and symbols, modern AI learns patterns from massive amounts of data, mimicking the biological structure of the brain's neurons. However, even these advanced neural networks still rely on the foundational premise of computationalism: that intelligence is ultimately a form of information processing that can be replicated in a machine.