Members

Uljana Feest

Technische Universität Berlin

Email: feest(at)mpiwg-berlin.mpg.de

Homepage: http://www.mpiwg-berlin.mpg.de/members/feest

Exploring Implicit Memory: On the Interrelation Between Operationalizations, Concept Formation, and Experimental Artifacts

The last few decades have seen the emergence of a new research field within experimental cognitive psychology and cognitive neuropsychology, dealing with a cluster of phenomena that are variously described as "implicit memory", "implicit learning", or "procedural knowledge". Hence, the emergence of this research field went hand in hand with the formation of a cluster of concepts. In Uljana Feest's project, special emphasis is placed on the concept of "implicit memory" and the experimental paradigms that are associated with the investigation of the phenomenon in question. The project is guided by the following two overarching theses: (1) The formation of a psychological concept (such as "implicit memory") is closely connected to the production of scientific knowledge claims about the (purported) phenomenon and the domain it is thought to be located in (e.g., the brain, the cognitive system). (2) There is a dynamic relationship between the formation of a concept and formation of the experimental systems in which such knowledge claims are generated. The notion of "scientific knowledge claims" is here understood to mean "beliefs that scientists currently take to be justified and capable of being true or false".

The approach underlying this project is situated within the context of recent writings by epistemologist of experimentation, who have argued that while experimentation is not as theory-dominated as traditional accounts would have it, it is also not as arbitrary or a-rational as some alternative accounts suggest. Now, given this de-emphasis of the role of theory in scientific experiments, three questions arise: (i) What is the status of scientific beliefs about a phenomenon if they are not, in any strong sense, embedded in, or derived from, a theory of the phenomenon? (ii) What is the rationale by which such beliefs might be considered justified by the outcome of an experiment? (iii) What is the sense in which experimentally confirmed beliefs are capable of (and can be classified as) being erroneous ? This project aims to develop a framework that systematically links these three issues.

The focus of the project is on the formation of concepts of phenomena about which little is known yet (in this case, the purported phenomenon of implicit memory). As scientists form more and more beliefs about the phenomenon, the concept of the phenomenon gets constructed and refined. Such belief formation is here referred to as a "process of discovery". The experiments that are conducted in this process are (in some sense) "exploratory" experiments. It is argued that in order to conduct such experiments at all, scientists have to rely on "operational definitions". Such definitions tentatively specify particular experimental paradigms as producing data, which are treated as indicative of the phenomenon in question (though it may later turn out that they are not indicative of the phenomenon, or that the phenomenon doesn't even exist). It is argued that experimental data become relevant to a given belief by virtue of the fact that experiments are frequently specifically designed in order to investigate a particular question. Feest refers to this process of translating a question into an experimental design as an "operationalization" of the question. Such an operationalization not only has to presuppose an operational definition for a concept of the purported phenomenon under investigation, but it also has to presuppose that all factors have been controlled for that might conceivably have a distorting effect on the data. Any given belief can only be justified by experimental data relative to those two types of background assumptions. Hence, the more reason scientists have to trust in the soundness of the background assumptions, the more reason they have to trust that their beliefs are not only justified but also true. Now, scientists usually don't refer to their beliefs as "true". They do, however, have an understanding of conceivable reasons for why a belief might be false. In this project, particular attention is paid to the way in which an erroneous belief might be due to an experimental artifact. To this end, an analysis of the notion of an "experimental artifact" is proposed, according to which experimental data are classified as artifacts if they were previously seen as confirming (justifying) a given belief, but are now recognized as having been caused by some processes other than those than they were previously thought to indicate.

With respect to the above three questions, this means the following: (i) while not derived from a theory, beliefs about a novel phenomenon are tested relative to prior conceptual assumptions (operational definitions) that underlie the experimental test of a research questions, (ii) beliefs are confirmed (and thereby justified) by data relative to such prior assumptions, and (iii) confirmed beliefs may turn out to be false if the data turn out to be artifacts (i.e., if prior assumptions turn out to be false).

Imprint