PAPER 1:
The Controversy between G. E. Mller and Wilhelm Wundt over the
proper measurement of reaction time.
Edward J. Haupt
Montclair State University
PAPER 2:
Perera, T. B. (1999) Reaction timing instrumentation
Paper presented in a symposium at the 1999 meeting of the
Eastern Psychological Association in Providence, RI. chaired
by Dr. Haupt and entitled:
The Muller - Wundt Controversy over
the Measurement of Reaction Time.
TEXT OF PAPER 1:
Revised version of paper presented at Eastern Psychological
Association Symposium, THE MšLLER-WUNDT CONTROVERSY OVER THE
MEASUREMENT OF REACTION TIME, Edward J. Haupt, proposer,
April 17, 1999. DO NOT QUOTE WITHOUT PERMISSION
Copyright (c), E. J. Haupt.
The Controversy between G. E. Mller and Wilhelm Wundt over the
proper measurement of reaction time.
Edward J. Haupt
Montclair State University
Table of Contents
Wundt's Program and the Importance of RT . . . . . . . . . . . 3
Means of Measuring RT as well as a little history. . . . . . . 4
Calibration of the chronoscope . . . . . . . . . . . . . . . . 6
A fundamental problem--Rate of Hipp spring oscillation. . 6
A fundamental problem--equalizing onset and offset
times. . . . . . . . . . . . . . . . . . . . . . . . 6
Wundt and Calibration. . . . . . . . . . . . . . . . . . . . . 7
Cattell's Practices . . . . . . . . . . . . . . . . . . . 8
Mnsterberg and RT. . . . . . . . . . . . . . . . . . . . 9
Mller's Critiques. . . . . . . . . . . . . . . . . . . . 9
Klpe and Kirschmann's response . . . . . . . . . . . . . 10
Edgell's Analysis. . . . . . . . . . . . . . . . . . . . . . . 11
Summary. . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
References . . . . . . . . . . . . . . . . . . . . . . . . . . 13
The Controversy between G. E. Mller and Wilhelm Wundt over the
proper measurement of reaction time.
Wilhelm Wundt is often taken as a paradigm of the precision
which accompanied the founding of the discipline of experimental
psychology. Much of this paper will question his use of reaction
time measurement to support that claim. First, I will examine the
centrality of reaction time for Wundt's program of psychology.
Then I will describe various types of reaction time measurement
in use in the 1870s and 1880s and the calibration problems which
arise from such devices. Then, from a variety of sources, I will
consider whether Wundt should be described as prescribing good
measurement practices. Finally I will discuss an extensive
contemporary report by the British psychologist, Beatrice Edgell,
which precisely evaluates the nature of reaction time measurement
shortly after the turn of the century. Edgell's report makes it
possible to understand how the goal, 1 ms accuracy, could be
achieved in an adequately equipped laboratory. The resulting
precision of measurement was a state of affairs that, for most
purposes, was maintained for 50 years and only ended with
improvements in the era of tube-based cascade timers and
transistors.
Wundt's Program and the Importance of RT
Wilhelm Wundt is generally recognized as promoting a program
for the study of "voluntary" psychological events. As such, two
features of each research procedure were of critical importance
to his studies. Each event must have a full opportunity for
introspection of the contents of the mind during the
psychological event (e.g., Benschop & Draaisma, 1999, p. 10) and
each time to respond must be measurable, thus permitting an index
for such mental processes. It is primarily with the latter
problem that I concern myself here.
Wundt first presented a study which used reaction time in
1862 (Wundt, 1862), while he was an Assistent for physiology
under the supervision of Hermann Helmholtz at the University of
Heidelberg. Thus, Wundt's interest cannot be traced to Donders'
experiments of 1867, as Benschop & Draaisma (1999, p. ) indicate.
What is more likely is that Wundt was making a metaphorical
extension of his supervisor's studies of nerve conduction time
from 1850-1852. In these studies, any mental process could be
seen as taking a specific neural route.
Danziger (1980, pp. 106-109) describes the importance of
reaction time for Wundt's "volitional" theories. Wundt started
from the assumption that when a reaction time was accomplished
with a muscular set, the time to respond was at a minimum. Wundt
(190x, Feldman, 1980) describes these as "... foreknowledge of
the object reacted to, and of the time of its appearance, and
absence of distraction. (p. 224)" Thus, several aspects of
experimental technique which Wundt considered requisites would be
considered bad technique at the present time. However, when the
reaction time was accomplished under a sensory set, the reaction
time could be varied as a result of the psychological processes
involved. Since the psychological processes were not established
experimentally, but were gotten directly from the introspection
by the participant in the experiment, these results are not,
properly speaking experimental studies, but correlational ones.
Wundt, thus, had a specifically voluntarist, non-
experimental program in mind when he started sponsoring
dissertations in 1879 in Leipzig. Fortunately, the first
dissertation of this series which was completed by Max Friedrich
(Friedrich, 1879/1979) has been translated by Peter Behrens.
This report shows rather clearly just what sort of a topic was
presented and how it was elaborated. In fact, one of the
surprising aspects of this study is that there is no discussion
of calibration of the Hipp chronoscope which was used. If
precision was the issue, this was an inauspicious beginning.
Means of Measuring RT as well as a little history
By the 1870s, there were a number of methods of measurement
for short intervals of time. For the very tiny intervals of nerve
transmission of the nerve controlling the gastrocnemius muscle of
the frog, Helmholtz had used the travel of a galvanometer needle
which he calibrated, following the ideas Pouillet (1844).
Wheatstone had created a chronoscope, and the improvements of
Hipp and others made it a standard device. The establishment of
the kymograph as a measuring device by Ludwig in 1847 had also
created the combination of the tuning fork and the kymograph
which was called the chronograph. However, there was little
systematic comparison of these methods in English until Beatrice
Edgell published a comprehensive article in the British Journal
of Psychology in 1906.
In 1839, Charles Wheatstone had invented the device to
which, as a result of improvements, the clockmaker Matthias Hipp
attached his name in 1843. This device was basically an
escapement-driven clock in which a tuned spring (actually, much
like a tuning fork) provided a precise driving speed of 500 and
then 1000 impulses per second through an escapement for a rotary
dial. Because of the necessary problems of accelerating the mass
of the dial to full speed through the weight driving it, the
tendency of the spring to sometimes double or halve its
frequencies, the difficulties of getting precise onsets and
offsets with the primitive relays of the time, and the varying
reliability of batteries, which continued to power electrical
switches until well after the turn of the century, chronoscopes,
as devices which had a visible dial were called, were problematic
for exact timing. The main advantage of the chronoscope was that
it had a visible dial from which a time value could be read, thus
eliminating the tedious and potentially unreliable counting of
cycles on a chronograph.
A second class of precise measuring devices were called
chronographs. Such devices made use of the recently improved
consistency of rotation speeds of the Marey and Ludwig-Baltzar
kymographs. The accurate regulation of the rotation speeds of
these kymographs was made possible by spring-regulated governors
in which a spring attached to a central shaft restricted the
outward travel of the movable arms of the governor, a device
which was labeled the Foucault governor. Such devices had become
the standard for accurate graphical recording of time signals by
the 1870s, and when the carbon-coated paper was removed from the
drum and fixed in a shellac bath, a permanent and publishable
record of a time signal.
These chronographs relied upon a tuning fork to provide an
accurate time base. Lest you believe that this device is yet
another Rube Goldberg contraption, let me make a case for the
accuracy of tuning forks. We all know that some piano tuners are
prodigiously accurate, and we can presume that similar paragons
staffed the tuning fork manufacturies of 19th century Europe.
However, any physics course will show you that tuning forks have
an easier potential for high accuracy of frequencies than many
other devices. This potential is found in the audible phenomenon
of beats, in which two tuning forks which are very slightly
different will produce a signal of varying loudness. The
frequency of this varying loudness is the difference in frequency
of the two forks, thus permitting easy adjustment of the erring
fork. Further (Turner, 1997, p. 33), Jules Lissajous, the French
physicist had perfected an optical standard for the calibration
of tuning forks and his work was the basis for the establishment
of the French standard for A (435 Hz). Thus, as long as the
batteries worked to move marking pens, a chronograph, the device
used by Donders (1865), had only the speed of the relays as a
measurement problem. Further, the time path of the relay could be
plotted simultaneously on the kymograph to determine calibration.
Since the speed of the tuning fork was absolute, the device
needed little calibration, but merely required a willingness to
count the number of tuning fork excursions, and a tuning fork of
200 or 256 Hz could easily accommodate a 1/1000 of a second
measurement. Wundt himself sponsored the creation of a larger and
more precise version of such an apparatus (Benschop & Draaisma,
1999, p. ), which these authors incorrectly suggest requires
calibration of the tuning fork (Benschop & Draaisma, p. 10).
A third set of devices depended on the precise measurement
of the velocity of a pendulum. Since the time for a pendulum to
travel from one side to another only depends on the length
between the edge from which it hangs and the center of its mass,
the length along a single swing of the pendulum could serve as a
time measure for periods of a fraction of a second. Since a
pendulum of 1 m has a period of 2 sec, a pendulum of manageable
size has the advantage that it could measure a wide variety of
typical human reaction times. The difficulty came in stopping the
pendulum precisely and thus measuring the location at which the
human reaction took place.
Pendulums were adapted for time measurement in other ways as
well. The American psychologist Sanford had devised a modified
"vernier" pendulum in which two pendula could be set to slightly
different lengths and an accurate difference of .02 sec per swing
could be found. There were also devices based on pendulums which
could provide accurate time bases for calibration of chronoscopes
or for timing stimulus presentation durations in other
experiments.
A fourth type of measurement was carried out by Helmholtz in
1850. Since, as a junior member of the K”nigsberg medical
faculty, he did not have funds to equip a physiology lab, he
resorted to a galvanometer excursion to measure short times in a
manner similar to Pouillet (1844). This, however, seems best
suited for very short intervals of a few microseconds.
Thus, in spite of the availability of alternatives, the
Wheatstone-Hipp chronoscope became a signature feature of the
laboratory of Wilhelm Wundt in Leipzig. This device, however, was
only one of several means for the precise measurement of time;
and it was a far more problematic device than has been portrayed.
Calibration of the chronoscope
The calibration of the Hipp chronoscope involved two
fundamental sources of error. These were problems created by the
spring which controlled the escapement speed and the onset and
offset times provided by the simple electrical relays which were
used to engage and disengage the clutch which caused the dial to
rotate.
A fundamental problem--Rate of Hipp spring oscillation
The spring which controlled the escapement frequency
presented two sorts of problems. One, which seems not to have
been much of a practical problem, was the tendency of the spring
to oscillate at twice or half the nominal speed.
The second, however, largely unnoticed, provided a
consistent bias. The frequency of the springs was, apparently,
not exactly 500 or 1000 Hz. It was 512 or 1024 Hz. This frequency
presumably occurred because the springs were made by musical
instrument makers, and 512 was the C below the treble clef.
Musical instrument makers (remember those Swiss music boxes) were
presumably skilled at making devices which resonate at such
frequencies, and would have charged much more for special
frequencies. Thus, RTs had to be adjusted as 1.024 the measured
time. Tedious, but not unthinkable, and a mechanical multiplier
would do a good job or a table of equivalent times. By the turn
of the century, however, the rate of oscillation was in fact 1000
Hz and the springs could be adjusted to the correct frequency
(e.g., Edgell & Symes, 1908, p. 282).
A fundamental problem--equalizing onset and offset times
The most complicated problem of calibrating the chronoscope
was provided by the electrical switches or relays which
controlled the clutch which engaged the dial. If you remember, as
I do, what relays were like in the 1960s, you remember compact
devices, tightly constructed, which had permanent magnets and
whose largest problem was the tendency for the contacts to "burn"
or weld themselves together.
The equivalent device of the 1870s was not so neat. The
basic principal was the same. A dc current was allowed to flow
into a coil surrounding a movable magnet. When the current
reached sufficient strength, a strong enough field to move the
unmagnetized iron bar was created and the bar moved to force the
clutch plates together. The timing was ended when the bar moved
in the opposite direction. Until about 1900, the electrical
voltages were provided by batteries. Wundt's new laboratory had
dc voltage from the wall in experimental rooms in 1897 and Lillie
Martin wrote articles in the American Journal of Psychology about
the power supply for the new Stanford laboratories, which she
designed, in 1904 and 190?. At the same time, Edgell and Symes
(1906) did all their work in the London University Physiological
Laboratory with batteries as a current source.
Thus, for much of the time with which we are concerned,
batteries were a problem. Batteries sometimes worked, sometimes
didn't work, always gave less voltage as time went on, ...
Since different battery currents gave different onset and
offset times, the problem of calibration of the chronoscope was
exacerbated by the use of batteries. Wheatstone (Edgell & Symes,
1904, p. 59) clearly recognized that a large current to close the
clutch and a small current to release it provided the best
result, which was that the onset time and the offset time would
be approximately equal, thus providing a an unbiased result. Such
a requirement, however, does not accommodate batteries well,
since a battery which produces a large current runs down much
faster than one which produces a small current. The importance of
making these times equal was generally recognized (Cattell,
Edgell & Symes, 1904, p. 60; Klpe & Kirschmann cited in Edgell
and Symes, p. 61).
Wundt and Calibration
It would seem that Wundt had an imperative to make sure that
the Hipp chronoscopes which he had in his laboratory were well
calibrated. Yet, Wundt was not famous for the knowledge required
or such precision. In 1859? Wundt had sought to become Assistent
to Emil du Bois-Reymond, recently appointed to the chair of
physiology created in Berlin after the death of Johannes Mller
when the teaching of anatomy and physiology was divided. du Bois-
Reymond was an important physiologist of the 19th century. He was
the central connecting link that made himself, Brcke, Ludwig,
and Helmholtz into an international institution. He was, also, in
many ways the founder of the field of electrophysiology, an
editor of an important journal (Annalen der Physiologie), an
experimentalist determined to see that replications always
worked, successor to Alexander von Humboldt as president of the
natural science section of the Berlin Academy of Sciences, and,
along with Hermann Helmholtz, one of those who were invited by
the Emperor and the Crown Prince to discuss the latest results
from the world of science.
In an part of his long exchange of letters with Helmholtz
(H”rz & Wollgast, 1986, p. 230), to whom du Bois had recommended
Wundt in 1859, du Bois comments that Wundt's recent book on
physics for physicians (1868) had made a complete mess of
electricity and makes several other points about Wundt as a
laboratory person.
Etwas Schlechteres als die Darstellung der Elektrizit„t in
Wundts medizinischer Physik ist selten dagewesen.
Vielleicht ist die Politik das geeignete Feld fr ihn. As
unbegabt fr die experimentelle Richtung hab' ich ihn an dem
Tage erkannt, wo er in meinem Laboratorium ein Stck Holz an
nasse Pappe mit Siegellack anzukleben versuchte und wo er
meinen sch”nen L[on]dner Drillborh[er] verdarb, indem er ihn
mit der Hand drehte, weil er den Mechanismus bersah oder
nicht verstand, der ihn in Rotation versetzt.
(Something worse than the presentation of Electricity in
Wundt's Medical Physics is rarely to be seen. Perhaps
politics is the suitable field for him. I recognized him as
ungifted for the experimental direction, during his days in
my laboratory when he tried to use lacquer to stick a piece
of wood on wet cardboard and when he spoiled my pretty
London drill press, in which [process] he turned it by hand,
because he overlooked or didn't understand what placed it in
rotation.)
While private reputation among physiologists of Wundt's
reputation may not be the best place to look for his actual
calibration practices, we might look to the laboratory practices
in order to determine what sort of calibration was used. The
first published dissertation from Wundt's "laboratory", that by
Max Friedrich, has been translated by Peter Behrens and published
in 1979 in Psychological Research, the continuation of the once-
Gestalt journal, Psychologische Forschung. The remarkable aspect
of this paper is that there is no mention of calibration of the
Hipp chronoscope in this paper at all.
In fact, Wundt seems to go out of his way to confirm du
Bois-Reymond's poor opinion of him. Wundt (Edgell & Symes, 1906,
p. 60) recommended in early editions of his Grundzge that the
currents for onset and offset be equal--a result which was then
known to produce different onset and offset times.
It is only later, in 1887, that Wundt's journal publishes a
paper by his student, L. Lange, which offers a Controlhammer, or
hammer-like checking device (the German meaning of control) which
can be used to calibrate the chronoscope, but only for durations
of 160 ms, thus leaving much of the measurement region which
interests psychologists entirely out of the picture. It was also,
only at this time that Wundt's journal presented the expanded
Chronograph, which has claims to accuracy of 1/10,000 of a
second. A claim which does not seem to have been validated and
which is of suspect validity.
Cattell's Practices
Benschop & Draaisma (1999?) have summarized much of what can
be extracted from the John McKeen Cattell diaries so ably edited
by Mike Sokal. While other reports from the Leipzig laboratory up
to that time do not show much interest in calibration, Cattell's
letters and diaries show that he obsessively controlled his
personal and laboratory life in order to carry out his timing
experiments carefully.
The implicit contrast between Wundt and Cattell suggests
several possibilities. One possibility is simply that Wundt's
practices were appropriate and Cattell was simply out of his
mind, responding to the mood of the times to feign precision and
gain respectability. This certainly seems to be what Lincoln
Steffens thought when he visited Leipzig and met Cattell.
I think it more likely that Cattell, who generally seems to
have been acutely oriented to the value of apparatus, was
concerned about the appropriate use of his equipment, and using
the measurements he collected, attempted to bootstrap his
procedures into a more precise quantification of what he was
measuring. In the mid 1880s only a few buildings in the world had
yet been equipped with electric power. Both wet and dry cells
created frequent problems and were unreliable sources. Cattell
clearly responds to this lack of reliability. Cattell limited
himself and his co-student, Berger, to series of only 26 reaction
times before an enforced rest period. Anyone who knows laboratory
technique for reaction times will see this precautionary measure
as a powerful guarantee that there is good attention for each
response.
Mnsterberg and RT
In 1888?, Hugo Mnsterberg finished his dissertation under
Wilhelm Wundt and accepted an appointment as Privatdozent at the
smaller Baden university in Freiburg im Breisgau under the
sponsorship of Alois Riehl. There, even though he had no salary,
using his own funds, he set up an experimental psychology
laboratory in his own house. He then published four issues of
Beitr„ge der experimentelle Psychologie, which reported his own
studies as well as the studies of some of his students. It is
undoubtedly this initiative in creating experimental studies
which brought him to the attention of William James, who had
annotated copies of all four issues in his library.
These privately issued papers contain a number of
interesting aspects. Mnsterberg did several extended
discrimination reversal studies (switching a watch from one pants
pocket to the other; switching the inkwell from left to right on
the desk) which seem to be the first proactive transfer studies.
While the studies are interesting because of their topic, some of
Mnsterberg's ideas seem less attractive, e.g., he thought that
with progressive reversals, the discrimination became weaker.
Mller's Critiques
G. E. Mller could be somewhat of a scold. I have found
examples of criticism in archives which verge on the gratuitous.
In addition, Mller wrote at least two papers which he titled as
Berechtigung auf... (Correction to) and published in significant
German archival sources. One of these corrected John McKeen
Cattell for having the temerity to suggest that Fechner's
psychophysical law could be expressed as a square root function.
Cattell seems to have thought that submitting was better than
arguing.
Two of Mller's critiques were directed at Mnsterberg and
concerned the measurement of reaction times. The first (Mller,
1891) was a review of Mnsterberg's Beitr„ge while the second
(Mller, 1893) concentrated on what was wrong and labeled itself
a Berechtigung. Hale (1980, pp. 24-25) has a summary of Mller's
critique, which is, essentially, that even though Mnsterberg has
rejected Wundt's metaphysics, he is not a very good experimenter
when the design and interpretation of the experiments is
considered.
A large part of Mnsterberg's studies were reaction time
studies which followed in the tradition of Wilhelm Wundt, but
which did not espouse the same voluntarist core which was present
in Wundt's program. Mller (1893) criticized these for their
faulty measurement of reaction time. In particular, he criticized
the use of the Lange Controlhammer as a calibration device,
maintaining that the error would increase systematically for
intervals beyond the duration that the Controlhammer could create
(about 160 ms maximum), and thus, all the reaction time
measurements were inaccurate.
Klpe and Kirschmann's response
In 1893, Wundt published an extensive description of methods
of calibration of the Hipp chronoscope authored by his two
Assistenten at the time, Oswald Klpe and August Kirschmann.
While Kirschmann was soon to go off to Toronto and create a
laboratory there, Klpe must be counted as the significant figure
for this topic, since he later (1902) served as Beatrice Edgell's
Doktorvater.
Klpe is usually portrayed as one of Wundt's loyal students,
in spite of the critiques of Wundtian psychology which came from
his laboratory during the "Imageless thought" controversy. I
think the question of Klpe's affiliation is more problematic
than it would thus seem. Klpe studied with G. E. Mller for
three semesters before Klpe became Wundt's first paid Assistent.
Mller and Klpe were in relatively frequent contact by letter
during the time Klpe was at Leipzig, from 1888 to 1894, when
Klpe became Ordinarius for philosophy in Wrzburg. One of
Mller's letters says that Klpe may be disappointed in Wundt
because of Wundt's "inexactitude." In the years of new century,
Klpe was present at the founding (1903) and a board member at
the first meeting (1904) of the Deutsche Gesellschaft fr
experimentelle Psychologie, which Mller headed for 23 years and
which Wundt never joined. Further, Mller students like Lillie
Martin frequently worked in Klpe's lab when they spent time in
Europe and Klpe students like Narziá Ach habilitated in Mller's
lab.
Edgell's Analysis
Beatrice Edgell was a remarkable early British woman
psychologist who was named Lecturer in Philosophy and thus became
head of the Department of Mental and Moral Sciences at Bedford
College, an all-female branch of London University on 1897. We
know less of her than we should. Nevertheless, she had strong
beginnings and made important contributions to British
psychology. She was the principal teacher of psychology at
Bedford College London from 1898 until her retirement in 1933.
Until her death in 1948, she continued to work from her
retirement home and to serve as an examiner in psychology for the
Royal College of Nursing, a post which she had occupied prior to
her retirement. She studied in Wrzburg with Oswald Klpe and
presumably met Narziá Ach during the academic year 1900-1901 and
completed a doctorate in 1902 with the title, Experimentation as
a border of psychology. She returned to England where she
established the laboratory at Bedford college (her first grant
for equipment and supplies was only $25), worked with prominent
British psychologists, among whom was Sir Frederic Bartlett. She
wrote (with F. Legge Symes) a major article on reaction timing.
For these articles, she listed her affiliation as the
"Physiological Laboratory of the University of London"). I think
this affiliation reflects Symes's position and was likely used
because only the Physiological Laboratory could provide the
equipment necessary for the studies they carried out. From 1911
to 1931, she made contributions to the Proceedings of the
Aristotelian Society that now would be regarded as essays on the
theory of mind (Valentine, 1999).
One of her significant achievements was the extended paper
in the second volume of the newly founded British Journal of
Psychology on the measurement of reaction times which, even
though it appeared partly in 1906 and partly in 1908. In this
article, Edgell and her collaborator, W. Legge Symes (later
Professor of Physiology at the Royal Veterinary College) describe
their measurement devices for showing the problems of the
Wheatstone-Hipp chronoscope as similar to those of Narziá Ach,
who provided an appendix about such problems in his 1905 book,
Ueber die Willenst„tigkeit. It should be pointed out that while
Ach was a doctoral student of Klpe, this book was the result of
Ach's Habilitation, which was carried out in G. E. Mller's
laboratory in G”ttingen, where Ach served as Assistent.
The main feature of the Edgell and Symes presentation is the
use of a light lever attached to the armature of the clutch which
engages the moving hand of the chronoscope. The position of the
armature can thus be traced on same kymograph drum which has a
tuning fork of 200 Hz providing a time base tracing (which means
that you must read the tracings to 1/5 of a cycle to get 1
millisecond accuracy). In addition, Edgell and Symes provide a
baseline which shows the position of the chronoscope armature
when the clutch is not engaged. In addition, they provided a mark
which shows when the clutch is engaged, as well as a line halfway
between these two which shows a putative make- or break-point for
the armature.
Edgell and Symes finally indicate that the parameters which
govern the time needed to make and break the chronoscope clutch
contact are different. In order to determine the time to make the
clutch contact, the spring tension which holds the clutch apart,
the current which is allowed to flow in the coil around the
magnet, and the electrical resistance of the circuit must be
known. For breaking the clutch contact, the mechanical inertia of
the magnet, dependent on its mass, and the remanent magnetism of
the iron bar must be known. Remanent (or remaining) magnetism of
the iron bar is a problem because, when the current in a coil
around a magnet is shut off, the bar retains some strength as a
magnet. The more times the magnet is used or the stronger the
current to make the contact, the stronger the remanent magnetism
becomes. In order to eliminate this problem of progressive
changes in the time to move the clutch apart, many chronoscopes
had been equipped with the ability to be operated by current in
either direction. In order to rapidly change the direction of the
dc current, a small switch, the Pohl'sche Wippe (Pohl's see-saw
switch) had been used.
In an added note, Edgell and Symes (1908) describe a new
chronoscope from Peyer and Faverger, the Neuchƒtel successor to
Hipp. She and Symes test the new device thoroughly, determining
that the frequency generator is inexact less than 1%. A failure
of the equipment then sends it back to Neuchƒtel, When it
returns, it is less exactly calibrated. However, they use methods
which they have previously described and soon the chronoscope is
even more accurate.
Summary
The accurate measurement of time intervals of less than 1
second became important when such intervals were used to infer
mental processes as they were by Hirsch, Wundt, and Donders. The
measurement of such time intervals by use of chronographs was not
problematic, but the use of chronoscopes was. The main error
appears to have been equalizing the time to make and break the
clutch contact. Equalizing these times was in fact a complicated
problem which depended on different sets of parameters for "make"
and "break" times. There are good reasons to suspect that Wundt's
own lack of electrical acumen as well as his failure to realize
these difficulties during much of his tenure at Leipzig produced
unresolvable errors in the times which were recorded. While not
all of the criticisms directed at Wundt were supportable, it
required several major papers (Klpe & Kirschmann, Mller &
Pilzecker, Ach, Edgell and Symes) before these difficulties were
fully resolved, relatively high dc voltages were readily
available, and Hipp's version of Wheatstone's chronoscope, as
manufactured by Hipp's successor, Peyer et Faverger, was a
reliable laboratory instrument. The main line of the studies
which demonstrated the necessary techniques to calibrate the
chronoscope occurred through G. E. Mller and his long-term ally,
Oswald Klpe, and their students.
References
Benschop, R., & Draaisma, D. (1999, in press). In pursuit of
precision. Annals of Science, ,?.
Danziger, K. (1980). Wundt's theory of behavior and volition
(pp. 1xx-1xx). In R. Rieber (Ed.), Wilhelm Wundt and the making
of a scientific psychology. New York: Plenum.
Edgell, B., & Symes, W. L. (1906). The Wheatstone-Hipp
chronoscope. A note. British Journal of Psychology, 2, 58-88.
Edgell, B., & Symes, W. L. (1908). The Wheatstone-Hipp
chronoscope. A second note. British Journal of Psychology, 2,
281-283.
Feldman, S. (1980). Wundt's psychology (pp. 2xx-2xx). In R.
Rieber (Ed.), Wilhelm Wundt and the making of a scientific
psychology. New York: Plenum.
Friedrich, M. (1980). An edited translation of the first
dissertation in experimental psychology by Max Friedrich at
Leipzig University in Germany (trans. P. Behrens). Psychological
Research, 42, 19-38.
Hale, M. (1980). Human science and social order: Hugo
Mnsterberg and the origins of applied psychology. Philadelphia:
Temple University.
H”rz, H., & Wollgast, S. (1986). Dokumente einer
Freundschaft. Briefwechsel zwischen Hermann von Helmholtz und
Emil du Bois-Reymond 1846-1894 [Documents of a friendship.
Correspondence between Hermann von Helmholtz and Emil du Bois-
Reymond 1846-1894). ( C. E. A. Kirsten, Ed.). Berlin: Akademie
Verlag.
Mller, G. E. (1891, Vol. I, 1. Juni). Review of
Mnsterberg. "Beitr„ge zur experimentellen Psychologie. Heft 1-3
[Contributions toward experimental psychology. Issues 1-3].
Goettingen gelehrte Anzeigen, 393-429.
Mller, G. E. (1893). Berechtigung zu Mnsterbergs Beitr„ge
zur experimentellen Psychologie [Correction to Mnsterberg's
contributions to experimental psychology]. Zeitschrift fur
Psychologie und Physiologie der Sinnesorgane, 4, 404-414.
In D. Cahan (Ed.), Explorations in 19th-century science.
Pouillet, M. (1844, 23. December). Note sur un moyen de
mesurer des intervalles de temps extrˆmement courts: Comme la
dur‚e du choc des corps ‚lastiques, celle du d‚bandement des
ressorts, de l'inflamation de la poudre, etc.; et sur un moyen
nouveau de comparer les intensit‚s des courants ‚lectriques,
soit permanents, soit instantan‚s [Note on a means to measure
extremely short intervals of time: Such as the duration of a blow
to an elastic body, that of releasing a spring, of inflammation
of (gun)powder, etc.; and on a new means to compare the
intensities of electrical currents, whether permanent or
instant]. Comptes Rendus des S‚ances de l'Acad‚mie des Sciences,
Tome XIX, S‚ance du 23 D‚cembre 1844. Institut de France.
Acad‚mie Royale des Sciences.
Turner, S. (1997). Demonstrating harmony: Some of the many
devices used to produces Lissajous curves before the
oscilloscope. Rittenhouse, 11, 33-51.
Valentine, E. (1999, June). Beatrice Edgell: An
appreciation. Paper presented to the annual meeting of the
British Psycholological Association. Available from: Elizabeth
Valentine, Department of Psychology, Royal Holloway College of
the University of London, Egham, Surrey, UNITED KINGDOM, UK-TW20
0EX.
-------------------------------------------------------------------------
Perera, T. B. (1999) Reaction timing instrumentation
Paper presented in a symposium at the 1999 meeting of the Eastern Psychological Association in Providence, RI. chaired by Dr. Haupt and entitled:
The Muller - Wundt Controversy over the Measurement of Reaction Time.
TEXT OF PAPER 2: PSYCHOLOGICAL REACTION TIMING 4-17-99 Copyright (c) 1999 Tom Perera Ph. D. Draft of paper for EPA presentation, April, 1999 OUTLINE: QUOTATION FROM THE UNIVERSITY OF BELGRADE COLLECTION BOOK Appropriate because of the current bombing of Belgrade BUILDING MUSEUMS ON THE INTERNET: Montclair State University Web Museum Barnard Web Museum Barnard Hipp THE HIPP CHORONOSCOPES Variations Hipp Reaction Time Setup Hipp Operation Tuning Fork Escapement Problems Start and Stop Solenoids Problems THE WET BATTERY INSTABILITY CALIBRATION OF THE HIPP Control Hammer Contact Pendulum IMPROVEMENTS AND CHANGES IN THE HIPP DESIGN Amerika Chronoscope D'Arsonoval Chronoscope Dunlap Chronoscope Springfield Timer CHRONO 'GRAPHS' Smoked-drum / tuning fork Phonograph Chronograph PENDULUM CHRONOSCOPES Barnard Pendulum Chronoscope Vernier Chronoscope GALVANOMETER CHRONOSCOPE THE HUMAN FACTOR Cattell with Wundt Cattell at Columbia THE COLUMBIA PSYCHOLOGY DEPARTMENT Keller's Introductory Classes Emphasis on wiring diagrams Knowledge of electronics Knowledge of Army Surplus Brenner's Shop German Precision Brenner Label FINAL SOLUTIONS TO THE TIMING PROBLEM Atomic Counter Quartz Crystal Oscillator Accurate Timer Expensive Hard to Interface Hunter Timers DISTRIBUTE BIBLIOGRAPHY.......... DEMONSTRATION OF: Battery Hipp Chronoscope Pendulum Chronoscope Vernier Chronoscope ******************************************************************** ******************************************************************** APPROXIMATE TEXT OF PAPER: ******************************************************************** PSYCHOLOGICAL REACTION TIMING 4-17-99 Copyright (c) 1999 Thomas Perera Ph. D. Draft of paper for EPA presentation, April 17, 1999 I would like to start out by reading an adaptation of the introduction to the magnificent book: Sense, Mind, and Measure: The Collection of Old Scientific Instruments of the laboratory for Experimental Psychology, University of Belgrade. by Aleksandear Kostic and Dejan Todorovic. I do this in recognition of the danger to both of them and to the extraordinary collection which they have put together in the face of the current bombing of their city. We can only hope that they and their collection will be spared. They write: "The Methods and spirit of Wundt's Laboratory were spread by many renowned European and American Psychologists. It is through an understanding of the methodology of Wundt's Laboratory that we can gain an insight into the development of Experimental Psychology itself." "The Psychological instruments which were the last word in technology have vanished from modern psychological laboratories. After multiple rennovations and house cleanings, most modern laboratories show no signs of any of the early apparatus." "I do regret that we have replaced that bountiful variety of switches, outlets, dials, wires, rods, tubes, supports, motors, threads, pens, those often imaginitive forms made of brass, wood, glass, rubber, plastic, and what-not for the boring monotony of computer screens and keyboards." "That is why we have worked so hard to preserve and restore the psychological apparatus in our museum." *********************************************************************** (THE CAPITALIZED HEADINGS ARE THE SLIDES WHICH I PRESENTED) *********************************************************************** MONTCLAIR WEB MUSEUM HIPP CHRONOSCOPE IN ZIMMERMAN CATALOG; Lacking the wealth of original instruments in the University of Belgrade collection, Dr. Haupt and I have set up an internet: Museum of the History of Psychological Instrumentation at Montclair State University which displays hundreds of pictures and descriptions of early apparatus. It is an evolving museum which currently shows the apparatus as displayed in the 1909 Zimmerman Catalog of Psychological Research Instruments. Zimmerman was the machinist and mechanic who made the apparatus for Wundt's laboratory and then went on to run the company which carried his name and which sold a wide variety of research apparatus. Dr. Haupt has laboriously translated the description of every piece of apparatus into English and we have made this resource available on the internet. The address of this and other internet sites as well as the titles of relevant books and articles are contained in a handout which is being circulated to all of you in attendance. BARNARD WEB HIPP I have been searching out and restoring some of the earliest research apparatus from the forgotten storerooms of Barnard College of Columbia University and I have made photographs and descriptions of the various instruments accessible to everyone on the internet as the "Barnard College History of Psychology Collection". BARNARD HIPP Barnard's most prized exhibit is this wonderful Hipp Chronoscope. It was used at Barnard College and I have brought it back to working condition by cleaning all the gears and bearings, and readjusting the mechanism. I am going to use slides to illustrate my talk about the Hipp and other timing devices and then I will demonstrate each piece of apparatus and give you a chance to see it in action and actually measure your reaction time with it. ** HIPP TYPES & OPERATION ******************************************************************** EUROPEAN HIPP TORONTO WEB HIPP: GLASS DOME, HIGH STAND WOODEN COVERED PENNSYLVANIA HIPP WOODEN COVERED MECHANISM There were many primarily cosmetic refinements of the Hipp Chronoscope and in its direct descendants as shown in these slides during it's 100 year production from the 1840's to the 1940's. HIPP SETUP: WORD STIMULUS, VOICE RESPONSE Here is a complete early reaction time experimental setup. A word stimulus is suddenly exposed and the subject responds by speaking loudly at this electromechanical voice key. The Hipp does the timing. HIPP FRONT VIEW Let me describe the operation of the Hipp and the intracacies of its mechanism. This front view shows the two dials. The upper one reads in milliseconds and the lower one increments one step for each (100ms) complete revolution of the upper dial. The mechanism is powered by a heavy weight which hangs down below the mechanism. START KEY START AND STOP LEVERS Some Hipp Chronoscopes were equipped with a key to start the mechanism as shown here. Others such as the Barnard Hipp which I will be demonstrating are started by simply pulling on a string. The string pulls the start lever and gives one of the gears a real kick-start which overcomes the resistance of the vibrating tuning fork escapement and starts it oscillating. CLOSE-UP OF TUNING FORK ESCAPEMENT: The tuning fork escapement had to oscillate at 1000 Hz and this required a very delicate adjustment as you will hear when I demonstrate it. All too easily, the frequency could shift to 500 Hz and the experimenter had to be constantly on guard to listen for this event and discard affected trials. This was one of the more easily managed sources of timing error with the Hipp Chronoscopes. START AND STOP SOLENOIDS CLUTCH CLOSE-UP PULL-IN / RELEASE TENSION SPRING FINE ADJUSTMENT DIAL PULL-IN / RELEASE KYMOGRAPHIC TRACINGS The major problem affecting accuracy came from the need to exactly equalize the rise and fall times of the start and stop solenoids. These solenoids moved this tiny bar from the stationary clutch face to the moving clutch face and back again. Adjustments were made by varying the return-spring tension using these fine calibrated settings. Keeping the Hipp calibrated was a tedious and difficult task. Due to variations in the above parameters and in the wet batteries used to power the solenoids, it was necessary to calibrate the Hipp every 20 or 30 trials. THE "CROW'S FOOT WET BATTERY The batteries consisted of Zinc and Copper electrodes immersed in a solution of copper sulfate. I have one such battery for you to inspect up here at the front of the room. Changes in room temperature, electrolyte concentration, and deposits on the electrodes caused the voltage and current output of the batteries to vary widely. ** CALIBRATION OF THE HIPP ************************************************************************* CONTROL HAMMER TORONTO CONTROL HAMMER OFF WEB The device which was used most frequently for calibrating the Hipp Chronoscope was the Control Hammer. The Control Hammer provided an accurate and repeatable set of contact-closures as a hammer literally fell past the electrical contacts. Due to limitations of size, it could only be used to calibrate time intervals up to 160 ms. CONTACT PENDULUM CALIBRATOR The Contact Pendulum was a second device used for calibrating chronoscopes and chronographs for longer time intervals than 160ms. It was capable of generating accurate signals separated by up to 2.5 seconds. Its operation was based on the constant oscillation period of a pendulum of known length. Dr. Galanter reports that at Pennsylvania, the pendulum itself had to be calibrated daily to compensate for changes in the room temperature which changed the length of the swinging arm. The Hipp continued to be used for 100 years but improvements were made and the improved instruments found their way into psychological research labs as their finances allowed. ** FURTHER EVOLUTION OF THE HIPP ********************************************************************** AMERIKA CHRONOSCOPE The first improvement was to replace the weight-driven mechanism with a wind-up spring driven mechanism while retaining the 1000 Hz tuning fork escapement. D'ARSONOVAL CHRONOSCOPE Next, a Foucault swinging weight governor was used to regulate the speed of the spring-driven mechanism. This produced a device called the D'Arsonoval Chronoscope. DUNLAP CHRONOSCOPE The next evolutionary step in chronoscope design was the Dunlap or John's Hopkins Chronoscope which used a synchronous electric motor which synchronized itself to the pulses produced by a tuning-fork oscillator. DUNLAP CHRONOSCOPE IN EXPERIMENTAL SETUP In this slide you can see a complete reaction time setup with the tuning fork that drove the Dunlap Chronoscope over on the right. Professor Galanter was responsible for such a Dunlap Chronoscope at Pennsylvania and reports that it sometimes took hundreds of spins to get the thing to synchronize with the tuning fork pulses and start to run. He likened it to trying to start a Model T Ford Automobile on a below-freezing day. SPRINGFIELD TIMER Finally, the Springfield Electric Company began producing the Standard Electric Timer which is still being sold. It was/is a Hipp-like mechanism with a constantly running synchronous electric motor which derived its extreme accuracy from the stability of the 60 Hz line voltages. INSIDE A SPRINGFIELD TIMER The electrically-operated clutch mechanism simply freed-up or stopped a wheel which was constantly trying to rotate as a result of the constantly-running electric motor. This mechanism was therefore very similar to the original Hipp mechanism. Electronic counters eventually replaced the electromechanical Springfield timers and brought us orders of magnitude better accuracy. ** CHRONO 'GRAPHS' ******************************************************************** PICTURE OF CHRONOGRAPH & HIPP ELECTRICAL TUNING FORK AND SINE WAVE TRACING ON SMOKED DRUM ChronoGRAPHS which wrote a tuning-fork produced stream of timing oscillations on a black smoked sheet of paper and imposed start and stop signals on the oscillations were introduced in the same time frame as the Hipp. They provided reliability and high degrees of accuracy but were extremely tedious to use because of the need to count each of the 1000 oscillations-per-second marks in a measured time interval. Because of this, 50, 100, and 200 Hz models were common. PHONOGRAPH CHRONOGRAPH For psychology departments that didn't have the money to buy the expensive Hipp Chronoscopes or Kymographic Chronographs, equipment manufacturers made devices that converted a standard phonograph into a chronograph. ** PENDULUM CHRONOSCOPES *********************************************************************** WEB BARNARD PENDULUM This is a very early reacton timer. Presenting the visual stimulus also releases a pendulum which swings along a calibrated scale until the subject stops it by pressing this bar which pinches the pendulum and brings it to a stop. BARNARD PENDULUM STIMULUS / TRIGGER BARNARD PENDULUM HAND - CALIBRATED SCALE BARNARD PENDULUM ADDED-ON SOLENOID ELECTROMAGNETIC STOP As you can see the pendulum apparatus was modified by the addition of electromagnets to allow electrical signals to instantaneously stop it's swing. WEB BARNARD VERNIER CHRONOSCOPE CLOSE-UP OF EXPERIMENTER'S AND SUBJECT'S FINGERS ON KEYS Another pendulum based mechanism was this vernier chronoscope. The two pendulums are set to slightly different lengths using a calibration rod which precisely positions each one at the proper length. One has an oscillation period of .80 sec and one has a period of .78 sec. The experimenter presses a button which starts the left one swinging and this is the stimulus to which the subject responds by pressing a button and releasing the second pendulum. The experimenter then counts the number of swings before they become synchronized and this number gives the number of 1/50'ths or .02's of a second that the two start times differ from each other. ** THE GALVANOMETER CHRONOSCOPE ********************************************************************** MIRROR GALVANOMETER CHRONOSCOPE Another type of chronoscope was made from a mirror galvanometer. It was appropriate only for very brief time intervals. An electrical voltage was impressed on a sensitive mirror galvanometer when the stimulus was presented. The galvanometer began to swing and to move a reflected beam of light. When the subject responded, the galvanometer decellerated and the position that the beam of light reached at the point of reversal of direction gave an accurate indication of the duration of the voltage from stimulus onset to response onset. ** THE HUMAN FACTOR IN REACTION TIME RESEARCH ****************************************************************** CATTELL AND WUNDT Finally, I would like to mention an all-but-forgotten aspect of improving accuracy in reaction time research. Clearly, the most accurate reaction timing devices are irrelevant if the subject daydreams, slouches, fidgets, and dozes during the experiment. Every fidget increases the variability of the individual reaction times. A highly trained and motivated subject, then, may be as important as the refinements in instrumentation accuracy. Cattell, shown standing on the right started out as a student of Wundt's in the 1880's and gradually set out on his own to bring accurate reaction time research to the United States, first at Pennsylvania in 1887 and then at Columbia in 1890. As a young man, he dedicated himself to becoming a 'professional' reaction time subject. In the same manner as an olympic athlete trains for the olympics, Cattell trained for his research sessions. He insisted that "inexperienced persons, children, or the insane be barred from participation." Cattell followed a strict and repetitive personal schedule from day to day consisting of constant exercise, personal discipline, and planned repetitiveness. PORTRAIT OF A MUCH OLDER CATTELL In this portrait which hangs in the Psychology Department at Columbia, you can see a man who has dedicated his life to the pursuit of accuracy and consistancy in research. ** THE TRADITION OF TECHNICAL PRECISION AT COLUMBIA ********************************************************************** KELLER'S INTRODUCTORY CLASS At Columbia, I was recruited from Fred Keller's Introductory Psychology Class to work as a lab assistant as soon as they discovered that I had started out as an Electrical Engineer and that I was a ham radio operator who knew how to buy and use Army Surplus Electronics. These electronic counters could be bought for a small fraction of their original cost either directly from the Government or from Radio Row in New York City, a dilapidated series of surplus stores which, sadly, were torn down to allow the construction of the world trade center. In those days, one's skills in electronics, mechanical construction, draftsmanship, and photography were as revered as one's knowledge of psychology. Notice that the left side of the blackboard is filled with cumulative record graphs and the right side has a complex electrical wiring diagram... At Columbia we had a wonderfully-equipped machine shop which was presided over (Or maybe I should say RULED) by Bob Brenner, a German Machinist who had been chief engineer on a German U-boat during WW-2. He often regaled us with stories of, for instance, turning down the diameter of the propellor shaft while under weigh to allow the installation of undersize shaft seals and stop a leak. (These skills might account for the survival of his U-boat in a war where most were lost. He brought these skills and the associated Helmholtz, Muller tradition of "PRECISE" German workmanship to Columbia and made all of our research and instructional apparatus with near-zero tolerances. He even signed his work with an engraved label shown here. BRENNER LABEL Prone to screaming outbursts of rage at our ineptitude with 'HIS' lathes, he terrified both students and faculty alike into emulating his obsession with precision and accuracy. ** FINAL SOLUTIONS TO THE TIMING PROBLEM ******************************************************************* When I entered the field of psychology over 40 years ago, the issues surrounding accurate timing had finally been solved. Ten years earlier, the development of the Atomic Bomb for WW-2 had required the development of high speed electronic counting circuits capable of capturing and counting the brief and rapid emissions of radioactive particles at rates of up to 100,000 events per-second. DECADE COUNTER SYSTEM Still earlier, in the 1920's, radio engineers had discovered that quartz crystals could be made to oscillate at extremely stable frequencies of 100,000 cycles per second and higher. CRYSTAL OSCILLATOR When the 100,000 cycles-per-second crystal oscillator was fed into a high speed counter, it became a clock which counted time in super- accurate steps of 1/100,000 of a second. (A hundredth of a millisecond.) For measuring reaction times, all that remained was to turn on the clock when the stimulus was presented, and turn it off when the response occurred. ELECTRONIC DECADE TIMER PRINTER ATTACHED One problem with these new counter/timers was that they were extremely expensive. This is the first timer brought into the University of Pennsylvania Psychology Lab. Dr. Galanter recalls that it took a plea to the president of the University to buy it and its associated printer. Another problem was that psychologists had no training in electronics and found it extremely difficult to interface the electromechanical relay programming equipment they had been using with the delicate high rise-time high-impedance pulse requirements of the electronic circuits. LATER STYLE ELECTRONIC TIMER EARLY HUNTER TIMER LATER HUNTER KLOCK-COUNTER TIMER These devices eliminated the early problems with reaction timing and allowed precise and repeatable measurements. ******************************************************************* DISTRIBUTE THE BIBLIOGRAPHY AND WEB SITE LISTING DEMONSTRATION OF BATTERY DEMONSTGRATION OF HIPP CHRONOSCOPE DEMONSTRATION OF PENDULUM CHRONOSCOPE DEMONSTRATION OF VERNIER CHRONOSCOPE -------------------------------------------------------------------------
THE MUSEUM OF THE HISTORY OF REACTION TIME RESEARCH
This museum traces the history of reaction time research.
THE BARNARD COLLEGE HISTORY OF PSYCHOLOGY COLLECTION.
This collection consists of color photographs and descriptions
of major early psychological research instruments.
CONTACTS:
NOTE: Dr. Haupt died of cancer in 2001.
Edward J. Haupt Ph. D.
Department of Psychology
Montclair State University
Upper Montclair, NJ 07043 USA
Thomas B. Perera Ph. D.
Professor Emeritus
Department of Psychology
Montclair State University