Introspective first person reports constitute the only available sourc
e of information about consciousness. Nevertheless, in order to develo
p and standardize techniques of analysis of subjective reports which a
re relevant to consciousness research, it is necessary to fulfill four
requirements: (1) to demonstrate their reliability and relevance, (2)
to establish criteria for selecting or obtaining the most appropriate
reports, (3) to develop a system for detecting the items in the text
which are indicative of conscious processes and, finally, (4) to devel
op procedures to represent such items and their structure and dynamics
with the aid of suitable formal devices. A narratological method whic
h meets the first two requirements and criteria is advanced. Despite t
heir obvious limitations, introspective reports can be considered to y
ield relevant and reasonably reliable information about consciousness.
Since certain introspective reports seem to be more relevant and reli
able than others, the question arises of discerning the most reliable
reports. The modern novel in the lines developed by Proust and joyce m
ay constitute the best representation of consciousness available. Desp
ite the fact that two examples from these novels show the extraordinar
y capability of language to convey mental states, the interior monolog
ue or the psychonarration of the modern novel are not optimally suitab
le to infer actual streams of consciousness because they are simulatio
ns. In some monologues, journals, autobiographies, and soliloquies the
writer expresses conscious mental states directly from his/er awarene
ss eliminating to a large extent the communicative intent. Thus, these
narratives retain more authentic traces of experience and they become
the most adequate targets for analyses of conscious processes. Carefu
ly selected excerpts from the journals of Virginia Woolf, Anne Frank,
Miguel de Unamuno and others show these characteristics. Other relevan
t items are constituted by verbatim transcripts of psychotherapeutic o
r self-experiment sessions. In all cases it can be asserted that these
are the most faithful reports of conscious mental states, and they ar
e called ''phenomenological texts''. Once a suitable phenomenological
text is selected or obtained a method to analyze it is required. It is
proposed that the text can be treated with some of the procedures dev
eloped by quantitative ethology and which include as a central require
ment an inventory of categories and a system of attribution and sampli
ng. Nine mental category terms (sensation, perception, emotion, though
t, judgment, reasoning, image, recall, intention) were used to analyze
an excerpt from the journal of a patient studied by Pierre Janet. Car
efully selected first person reports constitute ''phenomenological tex
ts'' suitable as targets of analyses for consciousness dynamic structu
re and process.