Is cultural evolution similar to biological evolution? That is the question posed by the authors of an article published on 20 May in the Journal of the Royal Society Interface. Sergi Valverde, visiting lecturer at the Department of Experimental and Health Sciences (CEXS) at UPF and Ricard Solé, ICREA researcher, members of the Complex Systems Laboratory of the Institute of Evolutionary Biology, have studied the natural and almost biological evolution of programming languages ​​that have profoundly marked the social and technological developments in the last 60 years.

Although many of the ideas of evolutionary biology have been applied to the study of cultural changes, there is still much debate as to whether natural and cultural phenomena evolve in a similar way. One of the reasons for the persisting debate is the difficulty of comparing the two processes. Unlike natural ones, cultural phenomena do not have a "genome" as a measure of change, with the exception of natural language for which formal language researchers have designed a number of metrics that represent grammatical, phonetic, orthographic and semantic changes, among others. However, other cultural phenomena, such as technological innovation, are much harder to quantify and measure properly.

So, Valverde and Solé have established an evolutionary model based on influence relationships of programming languages. As the authors comment, "programming languages are our bridge for communicating with computers, and they represent the biggest innovation in the history of contemporary technology. It is an invisible technology that has led to the information revolution and that has totally redefined our lives and the very nature of technological change".

Deciphering the Tower of Babel of artificial codes

Programming languages define a universe of diverse solutions. Some specialize in image management or in complex sound processing, and others consult large databases. To understand the complexity of information technologies it should be borne in mind that "programming languages have reached a stage of immense creativity and constant experimentation with new ideas", Valverde and Solé reaffirm.

The almost biological evolution of programming languages means that the structures they build are increasingly complex, as can be seen in the figure. One of the challenges the authors consider in their article is "to adequately define evolutionary trees in the history of programming languages".

The method the authors put forward has been developed using public data from Wikipedia, with the aim of revealing which programming languages have influenced others: "We use a simple measure of similarity between the languages that make up the network, very similar to the recommendation algorithms used by Amazon. Although the method is quite simple, it produces results that are totally consistent with the current knowledge on technological development", say the authors.

The results of data analysis and the modelling of the network show that the evolution of programming languages is very uneven, marked by waves of innovation in which new languages are created from the mixture of structures and concepts defined in preceding languages. These waves indicate the presence of short episodes of technological innovation consistent with the evolution of punctuated equilibria defined by S.J. Gould and N. Eldredge (1972).

Towards more natural human-machine communication

In addition to modelling the evolution of programming languages, Valverde and Solé's model offers new ideas on future developments. An important observation is the strong influence of the Artificial Intelligence on contemporary languages, a field that has traditionally been linked to the realm of academia and research. This suggests that some of the first ideas in computing represent innovations that were well ahead of their time. The tendency to develop languages that will enable more flexible and natural human-machine communication is also being seen to be reinforced.

But the true merit of Valverde and Solé's model is that the simple idea of trees of influence can be extended to many other fields beyond the study of programming languages, offering a new theoretical framework to rigorously quantify cultural and technological evolution.

Reference Article: Sergi Valverde, Ricard Solé (2015), "Punctuated Equilibrium in the Large-Scale Evolution of Programming Languages", Journal of Royal Society Interface , 20 May.