Université Lyon 2
Domaines de spécialité
Daniel Gile
daniel.gile@laposte.net
Outline of lectures
************************************************
Part I: An introduction to science
WHY SCIENCE?
How do we explore the world
around us?
- Experiential knowledge
(Knowledge
that we acquire by direct experience)
- Inherited knowledge
(Knowledge
that we acquire by reading other people’s statements and by listening to them)
Limited by:
- Sensory limitations
(We
cannot see through a wall, hear all sounds, discriminate between all smells,
etc.)
- Cognitive limitations
(We
cannot take in and remember all the information perceived by our senses,
draw inferences from
much information, plan in advance a chess game, etc.)
- Emotional interference
Centuries
ago, thinkers (philosophers), acknowledging these limitations,
especially the first two, began to develop science as a means
to push them back as far as possible.
*****
Science is made up of:
- An approach
(an attitude, norms)
- Institutional
means to help enforce theses norms
- Physical and intangible tools
(in particular mathematical tools, classifications, theories,
checklists, etc.)
***
CONCEPTUAL NORMS OF SCIENCE
Science is supposed to be:
SYSTEMATIC
(leave
no stone unturned)
LOGICAL
(every
step in an inference must be the logical consequence of a fact or previous
inference)
CAUTIOUS
(everything
is checked and double-checked, scientists hesitate to make claims or make them
on an explicitely
tentative basis)
OBJECTIVE
(trying
to avoid personal bias, in particular through the use of multi-person
assessments
and procedures designed to hide some potential
bias triggers from persons – ex: the double-blind
principle)
COLLECTIVE
(all
science is based on the collective effort of the community, and every
individual scientific endeavor
relies heavily on previous work by other
scientists)
COMMUNICATIVE
(publishing
one’s work makes the collective norm effective)
CRITICAL
(criticism
is one of the main drivers of science. Previous work is scrutinized, strengths
become a source
of inspiration, weaknesses a source of
improvement and further studies)
EXPLICIT
(explicit
reporting of one’s studies is essential if other scientists are to be able to read
it, assess it and get inspiration from it)
(see
http://teacher.nsrl.rochester.edu/phy_labs/AppendixE/AppendixE.html
)
***
WRITING NORMS
- Morphology
(IMRD in empirical texts)
- Citations
- Style
These will be examined more
closely with examples later in the course.
***
OTHER COMPONENTS OF SCIENCE
INSTITUTIONAL ORGANIZATION
(Universities, laboratories,
research institutions, etc.)
+
RESEARCH TRAINING
(In both universities and
research institutions, with degrees)
+
CONFERENCES, SYMPOSIA, SEMINARS
+
PUBLICATIONS - CENTRAL
(Journals, proceedings)
+
TESTS FOR PROFESSIONAL PROMOTION
(Theses, dissertations and
post-doc)
+
PEER REVIEW SYSTEM
(When publishing)
***
INITIAL CLASSIFICATION OF SCIENCE
BASIC RESEARCH:
LEARNING ABOUT THE UNIVERSE
APPLIED RESEARCH:
CHANGING THE UNIVERSE
(Closer to technology, but
applies scientific norms)
www.lbl.gov/Education/ELSI/research-main.html
***
(UTOPIAN) RESEARCH CYCLE:
OBSERVATION
(Of the object / Of previous texts about the object)
↓
THEORY
↓
TEST
↓
NEW THEORY
↓
TEST
etc.
(Karl Popper’s view)
(see
http://phyun5.ucr.edu/~wudka/Physics7/Notes_www/node6.html
)
Reality differs, because of technical and sociological
factors:
-
Testing
is not always easy
(material
and human resources, variability, measurability…)
-
Results
of tests not always clear-cut
(variability)
-
Paradigms
and groups in power, scientific revolution
(Thomas Kuhn)
( http://en.wikipedia.org/wiki/Thomas_Kuhn
)
***
What is a theory in science?
Theory is not “the truth”,
or “reality”
Theory is a mental construct which must do two
things:
-
Explain
known facts
-
Predict
phenomena
A “strong” theory is one
with predictive power
A “weak” theory has
explanatory power, but no or little predictive power
Scientists recognize that
theories will eventually be replaced with better theories.
They therefore want theories
to be “falsifiable”, that is, to be formulated in such a way
that they can be proved weak or wrong so as to be
improved or replaced by better theoretical constructs.
***
CLASSIFICATION:
THEORETICAL
RESEARCH VS EMPIRICAL RESEARCH
THEORETICAL RESEARCH:
Revolves around concepts
Thinking about data and what
sense it makes
Developing tentative rules,
laws, models or other conceptual structures
which account for
observations
Developing ways of testing
these rules, models, theories…
Learning about existing
theories, comparing them with other theories,
developing new theories on that basis
….
***
Example: The Effort Models of Interpreting
-
DG
noticed in the field that interpreting was difficult
-
Many
errors and omissions even when no noticeable difficulty
-
Introspection:
sometimes felt “not heard” something
-
Sometimes
forgot something before being able to reword it
Tentative intuitive conclusion:
Perhaps problems are linked
not to linguistic or thematic difficulty per se,
but to limited availability of some form of
“mental energy” required for interpreting operations
Developing a model
a. Tentative structuring of interpreting on the basis
of introspection:
3 Efforts:
Listening and Analysis LA
Production P
Short-term memory M
b. Each requires “mental
energy” (processing capacity)
Total available “mental
energy” is finite
Problems due to saturation
Interpreters work close to
saturation (“tightrope hypothesis”)
c. Explanatory power:
- Check whether this model
explains problem triggers
(fast
speeches, dense speeches, enumerations, syntactically different languages,
multi-word names, strange accents, strange logic,
etc.)
d. Further theoretical development
- Carry-over effects:
Problems at a distance
- Predictable sentence
endings
- Simultaneous vs.
consecutive
- Consecutive Interpreting
- Simultaneous with texts
- Sight translation
(see
Gile, Daniel. 1995. Basic Concepts and
Models for Interpreter and Translator Training.
Amsterdam/Philadelphia: John
Benjamins
Zhong, Weihe. 2003.
Memory Training in Interpreting. Accurapid Journal
7:3
http://accurapid.com/journal/25interpret.htm
)
***
EMPIRICAL RESEARCH:
Revolves around facts
(But is at the service of
theories, in that it is used
either to generate theories or to test theories)
In exploratory observation
of natural phenomena
In testing theories by
examining facts
***
BASIC APPROACHES IN OBSERVATION AND TESTING
- NATURALISTIC
Study phenomena as they
occur naturally
- EXPERIMENTAL
Create “controlled”
situations to study phenomena
***
MAIN ADVANTAGE OF EXPERIMENTAL RESEARCH:
A basically convenient way
to highlight the influence of variables one is interested in and to eliminate
the influence of interfering variables.
Example: the efficiency of a translator training
method M
You are interested in
investigating the relative efficiency of a particular translation method M.
One way would be to choose a
sample of students in many schools, and compare the quality
of translations done by students who have
been trained with method M with the quality of
translations done by students who have been trained
using other methods.
However, the quality of such
translations may depend on many other factors, for instance
standards at admission in the relevant schools, the
personality of the teachers, or the relative
difficulty of texts by which the students’
performance is measured.
In naturalistic studies, the
existence of such interfering variable (“confounding variables”) is
Generally
unavoidable.
One way to do away with it
would be to set up “artificial conditions”, in which in the same school,
the same teacher would teach two classes, one
with method M, and one with a more traditional method,
and test all students with the same source
text to translate.
In such an experimental
setup, the potential interfering effect of differences in admission standards,
teacher
Personality and source-text
difficulty are cancelled because these 3 variables are controlled, in this case
held constant.
***
MAIN PROBLEM: VALIDITY
Is the phenomenon studied in
the controlled environment the same as in the field, as it occurs naturally?
If not, conclusions drawn on
the basis of the experiment cannot be generalized to real life.
***
MAJOR TOOL IN (MOSTLY)
EXPERIMENTAL RESEARCH
INFERENTIAL STATISTICS:
USING MATHEMATICS,
- HELP MAKE INFERENCES FROM
SAMPLE TO POPULATION
-HELP DECIDE WHETHER OBSERVED
DIFFERENCE IS DUE TO CHANCE ALONE OR NOT
(IF PROBABLY NOT : “SIGNIFICANT”)
***
EXPECTATIONS FROM INDIVIDUAL RESEARCHERS
1. COMPLY WITH
NORMS
2. INNOVATE
INNOVATION:
- NEW FACTS
- NEW IDEAS
- NEW METHODS
GENERALLY, NO NEED FOR MAJOR
INNOVATION
HOW TO FIND INNOVATION OPPORTUNITIES?
- THROUGH
NEW OBSERVATION
- NEW QUESTION
- NEW INVESTIGATION FIELD
- NEW EXPERIMENT
- THROUGH
CRITICAL
PREVIOUS WORK
- FIND LOOPHOLES
(IN DESIGN, COVERAGE,
METHOD…)
- FIND WEAKNESSES IN
RATIONALE
- FIND ALTERNATIVE
EXPLANATIONS
***
INDICATORS
In order to be able to
investigate phenomena with accuracy and reliability, science uses indicators.
Indicators are discrete
units, preferably measurable.
(temperature, number of
occurrences, intensity of electrical current, binary states…)
Indicators need to be:
Reliable
(if used repeatedly to
measure the same phenomenon, they should produce the same result)
Accurate
Sensitive
Selective
Relatively easy to access
and process
Not too expensive
***
CAN SCIENCE EXPLORE ALL AND ANY PHENOMENON AROUND US?
The scientific method can be
applied to look at any physical, social, conceptual, emotional or other object.
But how efficient will be?
In particular, three major
obstacles can severely limit its power as an exploratory tool:
1. High variability:
When there is too much
variability in a phenomenon, it is difficult to identify patterns with
certainty.
2. Difficulty in finding
good, reliable indicators:
How does one measure
scientifically sadness or the quality of a dancing performance?
3. Lack of resources:
***
IS SCIENCE “BETTER” THAN OTHER TYPES OF INVESTIGATION
OR REALITY?
LIKE ALL HUMANS, RESEARCHERS
ARE LIMITED:
- PERCEPTUALLY
- COGNITIVELY
- EMOTIONALLY
SCIENTIFIC TOOLS AND NORMS HELP
THEM OVERCOME SOME LIMITATIONS
but
- EMOTIONAL FACTORS:
- HIERARCHY, TESTS AND
INDIVIDUAL COMPETITION
- COMPETITION BETWEEN
INSTITUTIONS
+
- STRONG SOCIOLOGICAL
FACTORS (Kuhn)
+
- CAUTION SLOWS DOWN
SCIENCE,
so
NON-SCIENCE OFTEN FASTER
EVENTUALLY SCIENCE IS BETTER
DUE TO SELF-CORRECTION
MECHANISMS
At any time :
SCIENTIFIC KNOWLEDGE
- ABOUT ISSUES WITH A
LONG HISTORY OF
INVESTIGATION
IS PROBABLY MORE SOLID
- ABOUT ISSUES WITH A
SHORT HISTORY OF
INVESTIGATION
MAY BE MORE SOLID
AMOUNT OF SOLID KNOWLEDGE
ABOUT ISSUES :
OFTEN SMALLER THAN
NON-SCIENTIFIC KNOWLEDGE
*****
SCIENTIFIC TEXTS
1. Due to the communication norm, writing and
publishing texts is an integral part of any researcher's activity.
As mentioned before, it is necessary for
his/her career, including promotion and tenure.
2. In a researcher's career:
·
Term
papers (students’ work)
·
Thesis
(MA or similar)
·
Dissertations:
Doctoral, post-doctoral/ Habilitation
·
Papers
(mostly in journals, sometimes in collective volumes)
·
Books:
monographs/collective volumes
·
Collective
volumes: thematic, honorary, proceedings of conferences, textbooks
***
THE STANDARD STRUCTURE OF
SCIENTIFIC EMPIRICAL TEXTS - IMRD
Introduction:
Literature review, statement
of the issue to be tackled, generally statement of the objective of the study
Method (and Equipment, Materials
and Methods, etc.):
Explains exactly how the
study was carried out, so that readers can understand, and assess and possibly
replicate the study.
Results:
Only presents the results, without discussing
them
Discussion:
Discusses the results, with comments as to how
they show this or that, support this or that hypothesis, or do not.
Often similar to a
Conclusion section
But before and after this
standard structure, other elements are found in papers:
Before IMRD
- Title, name of author(s)
[team, who is the boss], institutional affiliation
- Abstract, often with
keywords
- Sometimes acknowledgments,
especially when funded (acknowledgments are sometimes at end).
After IMRD
- Conclusion
-
Bibliography, often called “References”
- Sometimes
footnotes
- Sometimes
corpus
(reference:
www.mines.edu/Academic/courses/lais/licm598/Mark.doc
)
***
Linguistic features of scientific texts
-
Impersonal style
(has to do
with the objectivity norm; this is being challenged as hypocritical by some)
- Precise, non-repetition stylistic rule much weaker than in more
literary writings
- Much lexical innovation : scientists invent words and give
existing words new meanings because they define new conceptual entities
- Explicit
- Citation conventions
- Reference conventions
Example
of a scientific text :
Experimental
demonstration of the tomatotopic organization in the Soprano
(Cantatrix sopranica L.)
In Georges Perec’s
Cantatrix sopranica L.,
Paris, Editions du Seuil, 1991
Also see
http://pauillac.inria.fr/~xleroy/stuff/tomato/tomato.html