Husserl repeatedly stressed that his phenomenological method was a non-naturalistic alternative to the limitations of a naturalistic account of consciousness. Since science is performermed by scientists with their own consciousnesses, there is something circular about a scientific account of consciousness -- it presupposes its object of study. However, just because phenomenology is a philosophical discipline doesn't mean it can't inform experimental science. Phenomenology can be seen as supporting science. Naturalizing phenomenology can mean two things. Either phenomenology is adapted to fit natural science, or the concept of nature is adapted to fit the phenomena of phenomenology. We will concern ourself solely with the former. 1. Formalization The first approach is to translate the results of both phenomenology and neuroscience into a common, formal language. Having such a precise and mathematical language would guarantee successful communication about mental phenomena. Marbach (1993) argues that devising such a language must be possibible, since to the extent that phenomenology can describe experience through language at all, such descriptions might as well be formalized. The approach he suggests is to start with methodically collected phenomenological descriptions, and then to proceed by establishing a controlled, intersubjective and formalized terminology, which is the usual approach in scientific contexts. Marbach's notation is not about the content of experience but about its form, so it shows how the structure of experience can relate or differ from other experience. As an example, "[PER]x" signifies a person remembering having perceived x. But memory is always a re-presentation, according to Husserl, which involves a belief about the perception having actually occurred in the past. So the notation "(REP p |- [PER])x" signifies that x is being remembered as having been perceived. This view on memory is backed up by neurological findings that show that some of the same structures are implicated in perception and memory. Another proposal in this vein is from Roy, Petitot, Pachoud and Varela (1999). They also suggest mathematizing both third-person experimental data, and first-person phenomenolical descriptions. As a side-note, it could be remarked that Husserl himself was a trained mathematician, and he claimed that first-person data can never be fully reduced to mathematical formula. The reply of Roy et al. (1999) is that this assessment was based on the mathematics of Husserl's time, and that the development of dynamical systems theory removes the ground from Husserl's objection. Dynamical systems theory is the part of mathematics that deals with systems on or over the edge of chaos and describes them using differential equations. Their proposal stresses the importance of taking the full embodied experience into account -- the direct coupling in the sensori-motor system means that the object of study is naturally suited for the tools of dynamical systems. 2. Neurophenomenology A different approach that doesn't just focus on scientific models but also on experimentation is neurophenomenology, as devised by Varela (1996). The main idea is that not just experimental scientists but also subjects should receive training in phenomenological methods (though not necessarily phenomenology itself) -- primarily the epoche and the phenomenological reduction. That such training does not need to be too much of a burden can be appreciated by comparing the work scientists perform to train animals for experiments. Also, the experiments by Lutz et al. (2002) show the approach in action. This work focussed on variability in measurements that are standardly dismissed as noise by other scientists. It is hypothesized that this noise is caused by spontaneous thoughts and distractions, hence they can be called `subjective parameters.' In preliminary experiments subjects were asked to describe their subjective experience during a task involving visual stimuli. Afterwards the language of these reports was formalized in order to use it in the main trials where the subjective experience was correlated with reaction times and EEG measures of brain activity. The phenomenological methods employed boil down to the epoche, focused description and intersubjective corroboration. In the end it can be said that the experimental protocol applies a phenomenological reduction. Indeed results show that the reports of subjects correlate with EEG measures just prior to the presentation of the stimulus, as well as conditioning the resulting reaction times. 3. Front-loading phenomenology A third view on phenomenologically informed science is to introduce phenomenology not into the interpretation of results or the training of subjects, but directly into experimental design. This amounts not to presupposing phenomenological results, but to actually putting them to the test. An important difference with neurophenomenology is that subjects can remain completely naive about phenomenology, which implies that front-loaded phenomenology can function complementarily to neurophenomonology. An example of front-loaded phenomenology is the research on the difference between the sense of agency (being responsible for an action) and the sense of ownership. For example, moving your leg out of a kneejerk reflex involves no sense of agency, but it does involve a sense of ownership (it is your own knee that moves). According to phenomenology, this distinction is a first-order phenomenal experience, which does not rely on introspection. This implies that neuroscience should look for it as part of primary motor control processes, and not as part of higher-order cognition. Experiments such as Chaminade and Decety (2002), Farrer and Frith (2002) and Farrer et al. (2003) confirm this phenomenological analysis. References: Marbach (1993) Roy et al (1999) Varela (1996) Lutz et al (2002) Chaminade and Decety (2002) Farrer and Frith (2002) Farrer et al. (2003)