Abstract
In his post-1936 writing, Alan Turing developed the idea of an oracle machine, which, unlike the a-machine, as he called it – a machine that automatically executed all steps of a process until the process was complete – would consult an ‘oracle’. Although he did not describe the oracle in any detail, it can be assumed that the oracle would randomly, without resorting to a sequence-locked, procedural logic – configure decisions that would, or might, change the course of the rest of the process (Turing 1939). This means that at the very birth of computation, a different notion of computing, which flirted with randomness, incomputability arising from machinic interoception (Chaitin 2005; Ernst 2018), and unknowability, was considered a possibility.
At about the same time Vladimir Lukyanov built a water computer in the Soviet Union. Lukyanov’s aim was to understand the spread of thermal mass in metals used in rail track building, particularly their behaviour in extreme weather conditions. He had noticed that water flow was analogous to the distribution of heat and could act as a visual indicator of the thermal process. Lukyanov subsequently constructed a room-size computer out of sheet metal and iron and added glass tubes and tin vessels to monitor the water pressure in the system as water flowed through the glass tubes – which represented stored memory. The water computer produced numerical outputs which determined the answers to the complex differential equations that could be effectively solved with the water computer.
Utilising several other examples from early computing, such as Ross Ashby’s 1940s robotic homeostat, and W.G. Walter’s 1950s machine called speculatrix (a small wheeled automaton), this paper investigates the kinetic and potential energies of machines designed to both deal with uncertainty and unknowability, and uncertainty arising from the random, un-programmed aspects of the machines. A case in point is what is usually referred to as machine speculatrix’s self-awareness: the automaton’s capacity to ‘perceive’ itself moving through space by recognising and recording the sources, emissions, frequencies and/or velocities of thermal, light, magnetic and other energies.
Discussing situated iterative learning, responsiveness to stimuli, and auto-correction of operative parameters, the paper addresses the bigger question of the relation of particular forms of energy to machinic Umwelts and ecologies. Key aspects of this question are: Can machines that have been designed to deal with uncertainty in an adaptive manner (rather than through yes-no decisions) have a similar level of ecological porosity (understood as openness to the environment) as organic existents? And, how might this change their Umwelt (understood as an entity’s perceptual milieu) and Umwelt-based interactions? (von Uexküll 1930).
At about the same time Vladimir Lukyanov built a water computer in the Soviet Union. Lukyanov’s aim was to understand the spread of thermal mass in metals used in rail track building, particularly their behaviour in extreme weather conditions. He had noticed that water flow was analogous to the distribution of heat and could act as a visual indicator of the thermal process. Lukyanov subsequently constructed a room-size computer out of sheet metal and iron and added glass tubes and tin vessels to monitor the water pressure in the system as water flowed through the glass tubes – which represented stored memory. The water computer produced numerical outputs which determined the answers to the complex differential equations that could be effectively solved with the water computer.
Utilising several other examples from early computing, such as Ross Ashby’s 1940s robotic homeostat, and W.G. Walter’s 1950s machine called speculatrix (a small wheeled automaton), this paper investigates the kinetic and potential energies of machines designed to both deal with uncertainty and unknowability, and uncertainty arising from the random, un-programmed aspects of the machines. A case in point is what is usually referred to as machine speculatrix’s self-awareness: the automaton’s capacity to ‘perceive’ itself moving through space by recognising and recording the sources, emissions, frequencies and/or velocities of thermal, light, magnetic and other energies.
Discussing situated iterative learning, responsiveness to stimuli, and auto-correction of operative parameters, the paper addresses the bigger question of the relation of particular forms of energy to machinic Umwelts and ecologies. Key aspects of this question are: Can machines that have been designed to deal with uncertainty in an adaptive manner (rather than through yes-no decisions) have a similar level of ecological porosity (understood as openness to the environment) as organic existents? And, how might this change their Umwelt (understood as an entity’s perceptual milieu) and Umwelt-based interactions? (von Uexküll 1930).
Original language | English |
---|---|
Publication status | Published - 19 Sept 2024 |
Event | Forum on Philosophy, Engineering and Technology 2024 - Karlsruhe Institute of Technology, Karlsruhe, Germany Duration: 17 Sept 2024 → 19 Sept 2024 https://fpet2024.org/ |
Conference
Conference | Forum on Philosophy, Engineering and Technology 2024 |
---|---|
Abbreviated title | fPET 2024 |
Country/Territory | Germany |
City | Karlsruhe |
Period | 17/09/24 → 19/09/24 |
Internet address |