Abstract
In the mid-1970s, Soviet researchers explored alternatives to the rigid, deterministic logic that characterized early computing. This paper by A.M. Tiro proposes a framework of "informal calculus," designed to bridge the gap between algorithmic precision and the goal-oriented, often ambiguous nature of human intuition. By formalizing operations such as "system coincidence" and "system summation" relative to a specific objective, Tiro outlines a method for artificial intelligence to handle non-deterministic decision-making.
Introduction
The early development of artificial intelligence was often stymied by the fundamental difference between human and machine reasoning. While computers operated on strict, deterministic rules, human decision-making is inherently heuristic, context-dependent, and flexible. A.M. Tiro, writing from Tallinn, U.S.S.R., argues that existing heuristic methods were too specialized, requiring a bespoke "apparatus" for every unique problem. To address this, Tiro proposed "Informal Calculus"—a general system for intellectual decisions that treats ambiguity not as a failure of logic, but as a necessary feature of intelligence.
The Framework of Informal Calculus
The core of Tiro’s proposal is the move away from the assumption that a logical operation must yield a single, fixed result. Instead, the outcome of an operation is determined by its relationship to a specific "Goal" ($K$).
The Role of the Goal ($K$)
In informal calculus, the Goal ($K$) acts as the primary filter for relevance. It dictates which elements of a system are significant and how they should be prioritized. Without a goal, data remains a collection of neutral facts; with a goal, data becomes a hierarchy of functional meanings.
System Coincidence and Summation
Tiro introduces two primary operations to replace standard boolean logic:
- System Coincidence: This operation determines if two notions, $[A]$ and $[B]$, are functionally equivalent based on the goal $K$. For instance, a "desk lamp" and a "standard lamp" might coincide if the goal is "lighting," but they diverge if the goal is "portability."
- System Summation: This involves joining two systems to create a new one with emergent properties. Tiro cites "Mathematical Economy" as a summation of mathematics and economics, guided by a specific scientific objective.
Ambiguity as an Intellectual Feature
A central tenet of Tiro’s work is that "non-determination"—the ability to produce multiple possible results—is the hallmark of an intellectual system. He illustrates this with a thought experiment regarding the expression "1 + 8":
- Under an arithmetic goal, the result is 9.
- Under a typographic goal (creating a two-figure number), the result is 18.
- Under a geometric or aesthetic goal (creating a symmetrical figure), the result might be 8 or a symbol resembling infinity.
By allowing the system to generate multiple candidates and then filtering them through the Goal ($K$), the AI mimics the human ability to find creative solutions that a rigid algorithm would overlook.
Structural Representation and Functional Hierarchy
To implement this calculus, Tiro suggests representing systems as sequences of "numerals" (elements and links). These sequences are governed by a functional hierarchy: elements with the highest functional importance to the goal must appear first. A valid representation is one where the removal of lower-functioning elements does not destroy the system's ability to meet the goal, reflecting a robust understanding of core vs. peripheral data.
Practical Application: Industrial Production
While largely theoretical, the paper describes a successful implementation of the "system coincidence" operation on a standard computer. The model was applied to a real-world industrial problem: determining the "normal state of production." Despite having incomplete information, the system was able to assess whether a factory's production status was "normal" by algorithmically weighing a wide range of factors—much like a human foreman would assess a factory floor through experience and intuition.
Limitations and Conclusion
Tiro’s work was ahead of its time but faced significant technical hurdles. The author acknowledged that searching for "coincidental systems" required scanning the entire capacity of the computer's memory, a precursor to the "combinatorial explosion" problem that would later challenge symbolic AI. Furthermore, the high level of abstraction left many questions regarding the specific data structures required for implementation.
Nevertheless, "The Use of Informal Calculus" remains a significant historical document. It represents an early attempt to formalize the "subconscious" processes of human reasoning, suggesting that the path to true artificial intelligence lies not in more rigid rules, but in a structured embrace of ambiguity and goal-oriented flexibility.
Summary updated