Automata

John Clark sometimes he called the Eureka simply ‘The Automaton’, and it was one among many such devices that captured the popular imagination in the eighteenth and nineteenth centuries. The Eureka belongs to a well-established tradition of mechanizing both physical and mental tasks. There were ‘ingenious’ spectacles and curiosities, which included inventions such as Jacques Vaucanson’s Flute-player, exhibited in Paris in the 1730s; Kempelen’s automaton chess-player from the 1760s; the 1774 Jaquet-Droz androids, who, respectively, played music, drew, and wrote; and the Maillardet brothers’ Magicians, from the early nineteenth century, whose ‘tricks’ included answering a series of pre-determined questions. The mid-nineteenth century offered similar diversions: alongside the Eureka were Professor Faber’s Euphonia, a speaking machine (also exhibited at the Egyptian Hall) and Van Noorden’s Polyharmonicon, a machine that composed polkas.

Some of these devices, such as Kempelen’s chess-player, belonged to what the renowned mechanical engineer Robert Willis (1800-75), deemed to be lesser classes of automata: namely ‘the compound’ and ‘the spurious’. These devices, though mechanical to a greater or lesser degree, owed a significant portion of their operation to human intervention, though frequently this fact was not disclosed to observers. The Euphonia, for example, was a ‘compound’ automaton in that it could speak by means of ‘communication . . . with human agency’—that is, by compounding or combining mechanical apparatus with the input of a keyboard player. Kempelen’s chess-player was properly ‘spurious’: it functioned ‘under the semblance only of mechanism’ and was in fact ‘wholly directed and controlled by a concealed human agent’. Willis was not alone in speculating that its complex calculations were performed not by mechanical means at all but by a person hiding within a secret compartment.

Clark’s Eureka, though it belonged to the same milieu of showroom spectacle, conformed legitimately to what Willis termed the ‘simple’ category of automata: ‘those insulated Automata whose movements result from mechanism alone; by the aid of which they perform certain actions, and continue them, so long as moving force is kept in an active state’. Once set in motion, Clark’s automaton versifier produced Latin hexameters, one after the other, until its motive force—a descending lead weight that powered a system of gears, pulleys, and flywheels—wound down.

In his The General History and Description of a Machine for Composing Hexameter Latin Verses (1848), Clark himself acknowledged the connection between his device and certain curious automata. Unacknowledged by Clark, however, was another class of automata, arguably more ‘useful’ in their workings and applications than the curiosities discussed immediately above, whose mode of operation was in keeping with the Eureka’s piecemeal manufacture of verses. Indeed, as characterized by Andrew Ure in his 1835 The Philosophy of Manufactures, the whole of Britain’s factory system was fundamentally ‘automatic’ in that it effected production by machines, ‘with little or no aid of the human hand’. Just as ‘cords, pulleys, toothed-wheels, nails, screws, levers, [and] inclined-planes’ allow the Euphonia or a Jaquet-Droz draughtsman to perform its diverting functions, so too do these and similar mechanical compilations make possible large-scale ‘organizations [such] as cotton-mills, flax-mills, silk-mills, woollen-mills, and certain engineering works’. The factory system was for Ure ‘a vast automaton, composed of various mechanical and intellectual organs, acting in uninterrupted concert for the production of a common object, all of them being subordinated to a self-regulated moving force’.

The division of labour that underpinned the factory-system-as-automaton had much in common with the workings of both the Eureka and other automata. It depended upon a segmentation of processes, breaking down complex mechanical operations into a series of simpler tasks that could be performed in isolation from one another. Karl Marx understood that the structural logic on which the ‘division of labour, that distinguishing principle of manufacture’, depended was a form of serial manufacture: ‘the isolation of the various stages of production and their independence of each other’. Along with the machine refinement and segmentation of tasks that the division of labour encouraged, came what Marx viewed as a curbing of human input: fewer skills were required from workers, who became merely machine operators, their functions simply an extension of the machines they served. As he observed memorably in his Economic and Philosophical Manuscripts (1844), the worker in the modern factory system was alienated from the production process, reduced to ‘the condition of a machine’.

Not all of Marx’s contemporaries took this view. The mathematician and engineer Charles Babbage (1791-1871), for example, ‘considered the division of labour a key element in reducing production costs’ because it promoted ‘the invention of machinery able to perform . . . simple activities’, thereby reducing the cost of human labour and presenting the possibility of freeing up workers from ‘the duller, more repetitive activities’ that industrial manufacture necessitated. By ‘breaking a complex labour process down into simple operations’, the division of labour on which the factory system was organized at once diminished the reliance on ‘less qualified workers’ and stimulated technological innovation. In Babbage’s estimation, the factory’s mode of operating did not turn human workers into automatons, as Marx feared; on the contrary, it spared them that fate, purposely deploying machines, specially designed to function in line with the division of labour’s segmentation of tasks into ‘elementary components’, to reduce human drudgery while at the same time improving productivity.

Babbage is probably best remembered for his work to extend these principles to include a ‘division of mental labour’, particularly through his design of calculating machines. Believing machines might help to relieve the tedium of ‘computers’—a term, in Victorian usage, that denoted not machines but human clerks whose job involved basic reckoning or computation—Babbage conceived of a ‘Difference Engine’ that would both ease the tedium of manual calculation. By rendering calculation as a ‘systematic method’ that assigns to discrete components of the apparatus values that can be synthesized, the automaton calculator—which Babbage never managed to construct in full during his own lifetime—promised a solution to time-consuming numerical calculation and the production of reliable tables of logarithms, which nineteenth-century modernity’s impetus toward standard measurement methods, particularly in relation to accurate marine-navigation and ordinance-survey data, demanded.

Babbage eventually abandoned the Difference Engine, instead devoting his attention to an even more ambitious project: the Analytical Engine. While the Difference Engine could perform a pre-determined program, offering correct calculations that would be both time-consuming and tedious for a human ‘computer’, not to mention liable to error, it was, in essence, not much more of a legitimate ‘thinking machine’ than a Jacquard Loom or a Jaquet-Droz android. Where Babbage’s device automated a numerical formula, the others automated a sequence of physical movements—all were pre-determined, and none of the machines could ‘think’ beyond the formula or program (whether a punch-card or notched drum) they were given by their respective inventors. Babbage’s Analytical Engine, by contrast, promised to do more than calculate according to a pre-determined formula or ‘program’. As Niran Abbas points out in Thinking Machines (2006), ‘the Difference Engine’s successor . . . was a machine that not only calculated, but also decided what formula to use in order to do so’. Thus, Babbage’s Analytical Engine appeared to deliver what a device such as Kempelen’s automaton chess-player could only imitate by ‘spurious’ means—that is, a mechanics of deliberative action and intelligent discrimination, an ability to think about and decide upon a course of action without recourse to regular human input, either in the form of manual manipulation or via a pre-established formula or program. For this reason, Babbage’s Analytical Engine is often regarded as an antecedent of modern artificial intelligence devices.

Whether by means of deception or legitimate mechanical innovation, both ‘diverting’ and ‘serious’ automata, as exemplified respectively by Kempelen’s and Babbage’s devices, focused contemporary questions about the relationship between mind and machine. This is also where John Clark positioned his Eureka verse machine, itself an amalgam of showroom curiosity and putatively earnest endeavour. On the one hand, Clark imagined his device alongside Kempelen’s chess-player as an automaton that can ‘calculate’ an ‘uncertainty’ and then correctly mechanize a response to it. Instead of the moves of chess, however, Clark proposed an automaton that accomplished ‘a far more complex work . . . in a mechanical formation of Latin hexameters . . .’ The Eureka, as described by Clark, was able to mechanize the discriminating powers of the human versifier—whether schoolboy or poetic prodigy—who possessed not only an awareness of ‘the several kinds of verse’, namely ‘the several measures or Feet’, but also an understanding how best to combine these prosodic values to form a ‘poetical composition’. On the other hand, and more in keeping with the calculating automata conceived by Babbage, Clark was clear that his automaton versifier was not to be understood as involving any sleight of hand or other ‘spurious’ mechanics: there was no human hidden inside it.

Nevertheless, there is a considerable element of showmanship in Clark’s estimation of his machine’s compositional abilities. Though the Eureka did, in fact, employ legitimate mechanical means to manufacture its hexameters, its meters were not as ‘calculated’ and discriminating as Clark suggested, gesturing in only a basic way toward the arithmetical processes of the Difference Engine, much less the analytical ‘intelligence’ of Babbage’s later venture. What may have appeared, particularly to the metrically uninitiated among Clark’s audiences, to be a mechanical feat involving skills of linguistic and prosodic deliberation—a knowledge of Latin vocabulary and syntax, an awareness of quantitative prosodic values, an ability to combine foot varieties to compose a metrically ‘correct’ hexameter line—was, in fact, a formulaic process where values were pre-determined and their combination invariable. While there was in the operation of the Eureka an element of randomness, aligning it to an extent with Babbage’s work and its legacy in the a modern discourse of machine intelligence, there was also something fundamentally mechanistic in its verse manufacture, a quality that aligned it more with the unthinking and segmented automatic labour of the factory floor that a critic of industrial capitalism such as Marx regarded as contributing to a mechanic’s alienation from the mode of production in which he or she was engaged.

Rather like the schoolboys whose daily regime of scholastic poetry consisted of manufacturing nonsense hexameters one after the other, Clark’s Eureka was an automaton versifier, a metre robot that had no real understanding of the metres it manufactured. Further, if the verse compositions of the schoolboy were deemed by many to be lacking in any real utility, then there might be, paradoxically, a supreme usefulness in Clark’s hexameter machine in that it could relieve them, in a way similar to that imagined by Babbage, from that notoriously tedious task: their Latin verse composition homework.