Whether machines can think has concerned many philosophers for centuries. But, whether they can be creative and originate things is another fascinating question. The Analytical Engine illustrates the historical development of the debate concerning computational creativity.
The partnership of artificial intelligence and history might seem unusual. After all, history appears to be concerned with the study of the past – not with the technology of the future.♦ In this article, I intend to demonstrate through a discussion of the views of an important mathematician, Lady Ada Lovelace, that the contrary can be true: historical research can be valuable to research in artificial intelligence (AI).
Approaching Lovelace’s view historically
Ada Lovelace was born in early nineteenth-century London. She was the child of poet Lord Byron and, unlikely enough, the anti-poetical mathematician Lady Annebella Byron. From childhood onwards, Lovelace was pushed by her mother to pursue a scientific career, while simultaneously being withheld from the poetical influences of her father.
Lovelace’s ideas are often interpreted as a defense of
the idea that machines cannot be creative
In June 1833, the 17-year-old Lovelace began working with Charles Babbage, at the time a well-known scientist and inventor. This event marked a critical point in Lovelace’s thinking. A calculating machine that Babbage was working on, the ‘Analytical Engine,’ inspired her to develop her own ideas about machine creativity, culminating in 1843 with the publication of an important article, ‘Sketch of the Analytical Engine invented by Charles Babbage with notes upon the memoir by the translator’.♣
In the philosophy of AI, Lovelace’s ideas are often interpreted as a defense of the idea that machines cannot be creative.♠ However, this interpretation of Lovelace’s work is mistaken. To understand the views she actually held, we need to apply a historical approach. So, rather than assessing Lovelace’s views at face-value, I will put on my historical spectacles and ask the following question: what was the historical context in which she formed her views? Once we get a grip on Lovelace’s contemporal circumstances, it becomes clear that she was concerned specifically with Babbage’s machine, instead of extrapolating to all machines.
Babbage developed the idea of the Analytical Engine in the 1830s as a way of automatizing mathematical labor. Its purpose was performing error-free calculation, without the need for constant human intervention.1 With it, Babbage sought to eliminate the erroneous and arduous labor that manual calculation had traditionally demanded. Today, one might call Babbage’s machine a calculator.
Babbage never managed to finish his machine. Nevertheless, the concept of his machine combined with the small functional models he produced were sufficient to bring about a significant change in the understanding of machines in British society. His machine did not merely promise to make calculations more accurate and efficient – it would also mechanize mathematics, a discipline that had previously been exclusive to the human mind.2
Although his contemporaries had doubts about the project, Babbage saw the mechanization of mathematics as a perfectly logical course. In his view, mathematical calculation differed little from physical labor. Essentially, both activities could be mechanically implemented by dividing labor into smaller units. In turn, these units would jointly produce the complexity demanded for the task – whether it was the production of steam, the chopping of wood, or the calculation of complex functions such as y = 2x + x + 41. In short, Babbage believed that the right amount of mechanical complexity would do the job of mechanizing mathematics.3
Servitude of the machine
Lovelace deeply disagreed with Babbage. According to her, practicing mathematics is not the same as performing the correct calculations. Mathematics is something more than this: it requires understanding. In order to master this science, one needs to grasp the principles on which it rests. So, mere rule-following is insufficient for genuinely practicing mathematics.4
Consequently, Lovelace’s considerations led to a different view on the abilities of Babbage’s Analytical Engine. In her article, she warns against a particular overestimation of the machine’s powers: the Analytical Engine has no pretensions whatever to originate anything. It can do whatever we know how to order it to perform.5 Lovelace claims that the Analytical Engine cannot originate something by itself because it can only follow orders. To originate something is to create something ‘out of the blue’, without thereby following instructions. The Analytical Engine, however, cannot originate anything, since it can only perform according to the way we design it.6
During the period in which Lovelace wrote her article, no strict boundaries distinguished ‘state-of-the-art’ sciences such as mathematics from more ‘imaginative’ disciplines, such as poetry. In fact, different sciences often intermingled, resulting in a fruitful interaction between ‘strict’ and ‘imaginative’ sciences. Accordingly, the word ‘science’ did not refer to clearly defined fields. Rather, it referred to any systematic study in general.7 Notably, for British scientists, poetry was the heart of science. Without it, no scientific enterprise seemed possible at all.8 In order to practice a science, one needed to grasp its core principles. For that, however, one needed the poetical ability of imagination. So, poetry was not only beneficial, but indeed vital to understanding any scientific discipline.9
It needed to look beyond the rules that limited its capacities.
But the Analytical Engine could never do this.
Lovelace’s ideas were rooted in this scientific tradition. Appropriately, she referred to her own scientific approach as ‘poetical science’: a mixture of matter-of-fact skills and imagination. For her, imagination was essential to understanding the basics of any science: it was the means of ingenuity, the ‘discovering faculty’ that “penetrates into the unseen worlds around us, the world of Science”.10 Without imagination, one could at most follow the rules of a science, but one would never properly understand it.
Precisely because imagination was assigned such a crucial role in understanding science, a machine such as Babbage’s Analytical Engine, which followed strict rules, could never truly understand mathematics, according to Lovelace.11 In Lovelace’s view, creativity or ‘origination’ required Babbage’s machine to solve or produce new, unforeseen mathematical problems, without being itself regulated by rules. It needed to look beyond the rules that limited its capacities, and to understand what it was doing and how this could be done differently. But the Analytical Engine could never do this.12
In sum, Lovelace’s view on Babbage’s Analytical Engine was indebted to the scientific context of middle nineteenth-century Britain. She wrote about a machine specific to her time, in a scientific context widely different from ours. If one fails to take this into account, as some researchers in AI have done, it might be wrongly assumed that Lovelace claimed that no machine could ever be creative.
One might turn Lovelace’s case into a reflection upon other historical periods, or perhaps even upon our own day. Machines can often surprise us in their capabilities, simply because we cannot predict what machines can or cannot do in the future. To claim that machines can never be creative on the basis of the machines that we witness, is little different from gambling. After all, history shows that expectations about what machines can or cannot do have often been outpaced by unexpected developments. Perhaps our ideas about machine creativity will turn out false, too.
Editor: Dimitri van Capelleveen
♣ Lovelace’s article was an extensively annotated translation of an article by Italian philosopher Luigi Menabrea. Notably, Lovelace’s article became thrice the size of Menabrea’s original.
♠ For one of the early proponents of this interpretation, see: Alan Turing, “Computing Machinery and Intelligence,” Mind 59, 236 (1950): 450.
1 Swade, D., The Difference Engine. Charles Babbage and the Quest to Build the First Computer (New York: Viking Penguin, 2001), 17.
2 Swade, D., The Difference Engine. Charles Babbage and the Quest to Build the First Computer, New York: Viking Penguin, 2001, 85.
3 Sussman, H., Victorian Technology. Invention, Innovation, and the Rise of the Machine, California: ABC-CLIO, LLC, 2009, 43.
4 Toole, B.A., Ada, The Enchantress of Numbers (Sausalito: Critical Connection, 1992), 78.
5 Menabrea, L., “Sketch of the Analytical Engine invented by Charles Babbage,” in Scientific Memoirs, translated and annotated by Ada Lovelace and edited by Richard Taylor, 666-731 (London: Richard and John E. Taylor, 1843), 725-726.
6 Menabrea, L., “Sketch of the Analytical Engine invented by Charles Babbage,” in Scientific Memoirs, translated and annotated by Ada Lovelace and edited by Richard Taylor, 666-731 (London: Richard and John E. Taylor, 1843), 725-726.
7 Levine, G., “Defining Knowledge: An Introduction,” in Victorian Science in Context, edited by Bernard Lightman (Chicago: The University of Chicago Press, 1997), 21.
8 Lehleiter, C., ed., Fact and Fiction. Literary and Scientific Cultures in Germany and Britain (Toronto: University of Toronto Press, 2016), 9-10.
9 Casaliggi, C. and Fermanis, P., Romanticism. A Literary and Cultural History, London: Routledge, 2016, 4.
10 Toole, B.A., Ada, The Enchantress of Numbers (Sausalito: Critical Connection, 1992), 79.
11 Menabrea, L., “Sketch of the Analytical Engine invented by Charles Babbage,” in Scientific Memoirs, translated and annotated by Ada Lovelace and edited by Richard Taylor, 666-731. (London: Richard and John E. Taylor, 1843), 725-726.
12 Toole, B.A., Ada, The Enchantress of Numbers (Sausalito: Critical Connection, 1992), 72.
Cover image: Drawing by Thom Hamer
Image 1: Woodcut after a drawing by Benjamin Herschel Babbage, https://commons.wikimedia.org/wiki/File:Difference_engine_plate_1853.jpg
Image 2: Babbage Difference Engine, https://commons.wikimedia.org/wiki/File:Babbage_Difference_Engine_(Counter_detail_-_front_side).jpg