Operationalization of Meme Selection Criteria: Methodologies to Empirically Test Memetic Predictions Klaas Chielens Francis Heylighen Ecco,Vrije Universiteit Brussel Ecco,Vrije Universiteit Brussel Pleinlaan 2, 1050 Brussels, Belgium Pleinlaan 2, 1050 Brussels, Belgium klaas@chielens.net fheyligh@vub.ac.be Abstract This paper reviews a number of recent approaches to put memetics to the test of quantitative meas- urability. The focus is on the selection criteria for the spreading of memes put forward by Heylighen (1997), which include utility, novelty, simplicity, coherence, authority and proselytism. The general hypothesis is that memes scoring higher on these criteria will survive longer and be more prevalent than others. This can be tested by checking which story elements best survive a chain of person-to- person transmissions ("Chinese whispers" game), by simulating the cognitive and social processes that determine this differential survival and spread, and by correlating the score on the selection cri- teria with the actual frequency with which a meme is encountered. In a pilot study using an Internet survey, this method was applied specifically to virus hoaxes, which can be seen as paradigmatic ex- amples of clearly delimited, self-reproducing messages. 1 Introduction In 1976 Dawkins coined the term ‘meme’ to denote the cultural equivalent of the biological gene, i.e. an information pattern that is being copied from person to person. Examples of memes are jokes, ideas, tra- ditions, rumors, fashions and chain letters. Each of these information systems spreads by means of communication from one to several carriers. Thus, a successful meme can be compared to a cultural virus that "infects" a growing group of hosts. Over the past decade, an increasing number of publications has been devoted to memetics (e.g. Blackmore, 2000 & Aunger, 2001), proposing explanations for phenomena from viral marketing to consciousness and religion. However, the memetic approach has been criti- cized by many authors (Aunger 2001). Two major shortcomings can be pointed out: 1) it is hard to define what exactly a meme is; 2) the theoretical statements of memetics are as yet too vague to be empirically verifiable or falsifiable (Edmonds, 2002). The present paper proposes a broad method- ology to address these problems. We will argue that: a) a memetic perspective can suggest concrete and non-trivial predictions; b) given a suitable memetic unit of investigation, these predictions can be tested empirically. This should establish a firm operational footing for memetics, allowing a comparison of the strengths and weaknesses of different models, and thus transforming memetics from a collection of suggestive hypotheses into a true scientific disci- pline. 2 Meme Selection Criteria The core idea of memetics is that the popularity or success of a meme is determined by natural selec- tion. At any moment, several memes are in competi- tion for the attention of potential hosts and only those memes will spread that are well-adapted to the socio-cultural environment formed by these hosts and the network of their interactions; the others will become extinct. This leads to the generic prediction that “fitter” (i.e. better adapted) memes will become more widespread than less fit ones. To operational- ize this as yet very abstract (and to some degree tautological) idea, concrete selection criteria need to be formulated that specify the degree to which a meme is adapted to its environment. Several authors have tried to formulate principles that govern the spread of information. For example, Dawkins (1976), generalizing from the characteris- tics of biological evolution, listed the following three characteristics for any successful replicator, and thus for a meme: copying-fidelity, fecundity (number of copies made per time unit), and longev- ity (duration that any copy will survive). Working from a viral marketing perspective, Godin (2002) introduced the concept of the velocity with which ‘Idea Viruses’ spread from person to person. The social psychologists Schaller, Conway & Tanchuk (2002) focused on the communicability of a cultural trait. However, these characterizations of memetic fitness remain very broad and vague: what is it that makes a meme more communicable, fecund, or faster in spreading? They therefore offer little guid- ance in making non-trivial predictions.