Perceptual andMotor Skills, 1992, 74, 515-541.

O Perceptual and Motor Skills 1992

AUDITORY SUBLIMINAL STIMULATION: A RE-EXAMINATION '

MICHAEL J. URBAN

Coeur d'Alene, ID

Sumntary.-Unconscious or subliminal perception has historically been a thorny is-

sue in psychology. I t has been the subject of debate and experimentation since the turn
of the century. While psychologists now agree that the phenomenon of visual sublimi-
nal stimulation is real, disagreement continues over the effects of such stimulation as
weU as to its existence in other sensory modalities, notably the auditory. The present
paper provides an overview of unresolved issues in auditory subliminal sumulacion
which explains much of the difficulty that has been encountered in experimental work
in this area. A context is proposed for considering the effects of auditory subliminal
stimulation and an overview of current investigations in this field is provided.

Subliminal perception has had an uneasy relationship within the field

of psychology for many decades. There have been persuasive arguments for
(Dixon, 1981) and against (Holender, 1986) perception without awareness.
The present consensus, shared even by its former opponents (Merlkle, 1988),
is that subliminal perception is a real phenomenon. Left to be resolved are
questions relating to what kind of effects can be obtained by this method
and whether or not the auditory system shares the same capabilities that
have now been acknowledged for the visual system. The primary focus of
this paper is an effort to answer the latter question, since this issue has
become somewhat contentious within the community of psychologists.
I n the domain of visual stimulation, Hardaway's (1990) meta-analysis of
the subliminal psychodynamic activation (SPA) method developed by Silver-
man led him to the following conclusion:
. . . quantitative analysis of SPA studies has revealed a moderate and reliable treatment effect
for the stimulus MOMMY AND I ARE ONE that generalizes across laboratories and subject
populations. Criticisms stating that there are a dearth of weU-designed studies in this literature
are dearly the result of incomplete and biased sampling (p. 190).

Hardaway goes on to point out a particularly key element of these studies

which is that:
Treatment effects emerge only under subliminal viewing conditions (emphasis added) and appear
to be mediated by the emotional significance of the experimental scimulus (p. 190).

H e is joined in this assessment by Bornstein (1970), who reached simi-

lar conclusions in a separate meta-analysis. Based on these comprehensive
overviews it seems that the lingering controversy regarding the authenticity
of visual subception, as well as the subliminal symbiotic activation technique

'Address correspondence to M. J. Urban, Ph.D., Box 1768, Coeur d'Alene, ID 83814-1768.

5 16 M. J. URBAN

of Silverman (1978), is moving toward resolution. For clinicians, this repre-

sents fertile new ground for clinical investigations utilizing subliminal
"probes," as has been recently suggested by Dixon and Henley (1991).
The same sanguine assessment cannot be made for the acceptance of au-
ditory subliminal stimuIation, particularly with respect to the self-help ver-
sion that is commercially available on audiocassette tapes. A dispassionate
assessment of auditory subliminal stimulation has been confounded by at
least two major issues. The first is the aggressive and largely unsubstantiated
marketing claims of the subliminal self-help industry, and the second is the
sensational media coverage of the Judas Priest trial, in which the defendants,
a "heavy metal" band, were accused of embedding subliminal messages in
their music. These embedded messages were alleged to have been instrumen-
tal in the suicide and suicide attempt, respectively, of two young men. As a
result of this trial, many of the unresolved issues both for and against sub-
liminal stimulation have now been read into court records, largely on the
basis of testimonies by the same individuals who argue these issues in profes-
sional journals (Vance vs Judas Priest, 1989).
Since the defendants in this litigation stand to lose millions of dollars in
damage claims, they are predictably trying many tactics to win this case.
Among these tactics is a well orchestrated media campaign to discredit au-
ditory subliminal perception simply for purposes of legal advantage. While
this may be brilliant legal strategy, it makes for decidedly poor science. This
insinuation of legal and financial issues into the arena of subliminal percep-
tion has combined to obscure a technology which may be of considerable
value to the practicing clinician. Since the principles of science do not flour-
ish within the win-at-any-cost world of adversarial litigation, psychologists
would d o well to take a more protective position in defending both the con-
cept of preconscious processing and its attendant technology.
The following- sections present a brief comparison of several methods of
auditory subliminal presentation, give a synopsis of the present technical
state of the art in production of subliminal audiotapes, propose a new con-
text for reconsidering the utility of subliminal audiotapes, and review recent
and pending publications which used commercialIy produced subliminal au-
diotapes in their research designs.
Methods of Auditory Subliminal Presentation
Historically (Stroh, Shaw, & Washburn, 1908), the most common meth-
od of auditory subliminal presentation has been a simple reduction of sound
amplitude to the point that subjects could not identify the presence or ab-
sence of the sound at better than chance accuracy. For our purposes we will
call this the 'auditory threshold' technique. Merikle (1988) has criticized this
method as providing only "perception in the absence of subjective confi-
dence" rather than perception in the absence of awareness. This argument
 AUDITORY SUBLIMINAL STIMULATION: A RE-EXAMINATION 517

finds its historical precedent in the debates on s i d a r issues in visual sublim-

inal perception between Erikson (1956, 1960) and Lazarus (Lazarus &
McCleary, 1951; Lazarus, 1956). It is an argument that is not likely to be
resolved anytime soon as it is rooted in epistemological considerations and
beliefs that are quite different among clinically oriented practitioners and sta-
tistically inclined empiricists.
Even without resolving these deeper issues it is difficult to assess the
comparative results of the available auditory SPA studies. Those studies that
did not use a masking sound for the stimulus often did not report at what
decibel (dB) level relative to threshold the stimulus was presented (Buchholz,
1968169). Some investigators (Ben-Hur, 1979; Chimera, 1987) report mini-
mal differences (-5 dB) in their stimulation levels relative to threshold, while
others report stimulus/threshold differences as high as -30 dB (Shifren,
1981). Since an 8.3-dB increase in the amplitude of sound will be perceived
as a doubling of loudness, it is clear that these studies are not comparable,
even when considered solely from a psychophysical perspective. Furthermore,
at -30 dB below subjective threshold, Shifren (1981) may indeed have sat-
isfied Merikle's (1788) criterion of an objective threshold or the point at
which no physiological response is possible.
Other investigators have presented stimulus messages masked by music
or other sound at levels described only as "beyond the level of conscious
awareness" (Bouchard, 1984) or qualified by statements indicating subjects
were unable to discover the existence of a verbal message (West, 1984;
Doche-Budzynski & Budzynslu, 1989). Finally, Kaser (1986) presented stim-
uli which he claimed were subliminal by virtue of having been speeded to
the point of being unintelligible when consciously perceived.
Quite apart from the fact that the above cited studies all reported dif-
ferent results and used different protocols, they are impossible to compare to
one another for technical and psychophysical considerations.
From a technical point of view differences as small as 10 dB in signal
presentation between studies will make comparisons impossible, particularly
when the method used is simply decreasing the stimulus volume below the
subjective threshold. At these levels sound energy is exceedingly small and
minimal differences in stimulus presentation level can easily exceed physio-
logical response capabilities. Thus the method of simply reducing stimulus
amplitude as described in several of these studies (Buchholz, 1968169; Ben-
Hur, 1979; Chimera, 1987; Shifren, 1981) is the least precise, probably the
most likely to produce null results, and certainly most vulnerable to criti-
cisms about partial cues or thresholds.
A second method consists of masking the signal with music or white
noise. This is typically accomplished through the simultaneous presentation
of speech and noise, which can be recorded onto magnetic tape by mixing
5 18 M . J. URBAN

the two sound tracks together. Ideally the peak amplitude of the message sig-
nal should never exceed the amplitude of the masking signal. The chief
drawback in this method is that the actual level [measured as the signal to
noise (SIN) ratio in decibels] of stimulus presentation can vary markedly as a
function of the voice and speech characteristics, is inconsistent, and can
itself have a range of as much as 30 dB. Also, when using this method, one
typicdy will set the peak level of the message in some subjective fashion, so
this method is also open to criticism with respect to partial cueing.
A third method of producing an auditory subliminal stimulus is to use a
subliminal processing device. This technique, which was originally intro-
duced by Becker, Corrigan, Elder, Tallant, and Goldstein (1965), uses an
analog circuit to maintain a consistent SIN ratio between the masking sound
and the stimulus signal. I t is an unappreciated but important fact that this
method is superior to the simple mixing of a stimulus with sound and funda-
mentally different from the auditory threshold technique. The key dfference
between this method and those discussed previously is that an auditory sub-
liminal stimulus which has been processed by a mixer that adjusts the level
of the speech signal relative to the level of the masking sound presents the
-

subliminal stimulus at a consistent level relative to the masking sound rather

than to the subjective threshold of hearing as is the case in the auditory
threshold technique. This distinction is quite important as it speaks to the
issue of retrievability of a subliminal stimulus. In the case of the subliminally
processed stimulus, the retained sound energy of the masked message is well
within the auditory range of the subject at aU times.
For purposes of illustration Iet us assume that speech in the frequency
range of 1 kHz has a peak amplitude of approximately 60 dB. I n this case a
subliminal mix with a SIN ratio of -25 dB still leaves the message with an
amplitude of 35 dB which would be both audible and intelligible in the
absence of the masking sound. Also, the technique of subliminal mixing is a
method for subliminal presentation of auditory stimuli that is consistent and
reproducible, in contrast to the previous two techniques. Subliminal mixing
also presents a very different problem to the auditory system and the brain,
which is one of signal extraction from noise as opposed to one of signal
detection as in the auditory threshold technique.
Finally, recent advances in computer technology, and particularly digital
signal processing (DSP) techniques, have made it possible to refine Becker,
et al.'s ideas in some major ways which have significantly improved upon
performance of so-called subliminal mixers. The application of DSP tech-
niques has numerous advantages, one of which is an improvement of about
10 dB in S/N ratio. While there are additional advantages, some of w h c h
are discussed below, it is clear that this is the preferred method for produc-
ing experimental or laboratory-grade auditory subliminal stimuli.
 AUDITORY SUBLIMINAL STIMULATION: A RE-EXAMINATION 519

Subliininal Audiotape as a Sev-help Modality

The argument against efficacy of subliminal audiotapes derives primari-
ly from a single study by Greenwald, Spangenberg, Pratkanis, and Eskenazi
(1991). This study seems far from giving the unimpeachable evidence that
opponents of auditory subliminal perception suggest. Even if considered in
the context of placebo effects, the conclusions of the authors based on their
description of this study seem to be somewhat premature.
Given the textbook design of the study and its thorough statistical anal-
ysis, a cursory reading can be quite persuasive; however, good design and
analysis cannot substitute for a failure in conceptualization (Bergin & Strupp,
1972) that in this case seems consequent to an incomplete understanding of
the technical issues attendant to the production of subliminal audiotapes, an
issue which is discussed in detail below. Of particular significance is the sig-
nal to noise (SIN) ratio between message and masking medium and a thor-
ough knowledge of the script content of the programs. As has been demon-
strated by Zenhausern, Ciaiola, and Pompo (1973) and Zenhausern and Han-
sen (1974) and more recently by others (Kotzk & Moller, 1990; Swingle,
1992), the S/N ratio is of crucial importance in demonstrating a reliable sub-
liminal effect. Thls is of particular importance in the Greenwald, et al.
(1991) study since the authors indicate that they duplicated tapes from cas-
settes purchased through mail order. There is little question that such a pro-
cedure would add enough noise to a second or third generation copy to cre-
ate interference with retrieval of information by subjects.
A second concern of some import is the observation that Greenwald,
et al. (1991) did not know the message content of the various tapes under
investigation. The relevance of this issue is illustrated by the following exam-
ples. "I have complete confidence in my abilities" and "I have complete
confidence in my ability to remember" are statements from commerciatly
avdable self-esteem and memory programs, respectively. One would be hard
pressed indeed to demonstrate a differential effect of these statements even
if they were delivered audibly. This tendency of commercial tapes to contain
what are best described as "global" affirmations seriously undermines the
putative ability of the Greenwald, et al. (1991) study to differentiate be-
tween products. Indeed, it appears that the results of this investigation sup-
port the notion of a broad-based improvement on all measures consequent to
the subjects' exposure to these global subliminal messages. This improvement
is interpreted, mistakenly in my view, as an dlusory placebo effect.
A third problem is the lack of a true control group, one using an audio-
tape without a subliminal message. Given the confounding characteristics of
the preceding observations, the latter hardly needs additional comment. Fi-
nally, it is worth noting that the lack of equivalent manufacturing standards
among tape vendors and, indeed, the questionable quality of some products
argue against generalizing from discrete samples.
520 M. J. URBAN

An equally plausible explanation that cannot be rejected on the basis

of their design is that the "across the board improvements in memory and
self-esteem" as reported by Greenwald, et af. (1991) are due to the previ-
ously reported general arousal phenomenon of subliminal stimulation (Zen-
hausern, et ol., 1973; Zenhausern & Hansen, 1974; Zajonc, 1980; Swingle,
1992). The fact that virtually all of their subjects showed improvements in
the target criteria and in their subjective assessment of memory and self-
esteem irrespective of condition assignment weakens the placebo-only claim
and adds credence to the possibihty of the additional effects noted above as
well as content-specific effects of the subliminal audiotapes.
Regarding the placebo issue, the literature on placebo effects indicates
that there are subpopulations of responders as well as nonresponders and
negative responders (Modestin, 1985; Moertel, Taylor, Roth, & Tyce, 1976;
Wolf & Pinsky, 1954). The Greenwald, et a(. (1991) study is strilclng in its
lack of both nonresponders and negative responders. The ubiquitously posi-
tive outcome enjoyed by their subjects suggests that there are other factors
operating than those suggested by the authors. It would be a rare placebo
indeed that was 100% effective with 100% of subjects. I am unaware of
any other "placebo" intervention that is so uniformly positive in effect and
outcome. While expectancy is without question a significant component in
the effects reported by Greenwald, et al. (1991), it should be noted that
expectancy is a factor of all behavioral interventions. For the reasons stated
above, the present study cannot and does not rule out effects directly conse-
quent to the delivery of subliminal auditory stimuli.
Finally, since consensus is lacking as to the level or complexity of analy-
sis that can be initiated through auditory subliminal stimulation, there is
some question as to the appropriateness of the instruments ut~lizedfor out-
come measurements in this study (e.g., is the Wechsler Memory Scale really
an appropriate test for the effect of subliminal auditory stimulation?).
The second bit of evidence that is often brought against the efficacy
of subliminal audiotapes draws on a study published by Merikle (1988), in
which he used a sonogram to "test" commercial tape products for evidence
of embedded messages. This paper achieved a great deal of undeserved no-
toriety primarily because so few people in academic psychology understand
what a sonogram is or what it does. Individual researchers who oppose the
concept of auditory subliminal perception described this study as "provoca-
tive" (Eich & Hyman, 1991). In truth it is equivocal at best, as sonograms,
which graph spectral components as a function of their signal strength, do
not necessarily have the capability to extract information that is masked at a
signal to noise (SIN) ratio in excess of -20 dB.
Only under ideal conditions can one recover the residue of voiced
signals from broad-band masking. Unpublished work in thls laboratory has
 AUDITORY SUBLIMINAL STIMULATION: A RE-EXAMINATION 52 1

shown it is possible to retrieve some residual signals in noisy environments

up to a SIN ratio of -25 dB. These results are entirely dependent on first
having an unmasked sonogram of the speech sample of interest and then pro-
gressively adding broad-band masking until there is no longer any evidence
of the original speech signal. An illustration of this procedure is presented in
Fig. l a through If below.2 Attempting to demonstrate, via a sonogram, the
presence or absence of speech signals that have been embedded in noise ac-
cording to the methods described above, without first having a sample of the
speech signal of interest, does not appear to be a viable technique. I n the ab-
sence of a speech template for comparison purposes it is difficult to distin-
guish speech from noise by way of a sonogram even at SIN levels as low as
-10 to -15 dB. Thus one can conclude that Merikle (1988) is mistaken in
his assertions with respect to the sonogram data. The many factors which can
affect the results of spectrographic analysis include SIN ratio, the particular
speech pattern of interest, spectral characteristics of the masking sound, rate
of speech, spectrographic analysis bandwidth, and even the contrast ratio
used for graphing the plots. None of these factors were considered by
Merikle (1788) in reaching his conclusion that:
If messages are actually embedded within the background sounds, then it should be possible, at
least occasionally, to identify the unique energy patterns associated with speech in the spectro-
grams produced From subliminal cassettes (p. 361).

This speculative argument was the basis of his subsequent analysis. The
fact that he did not produce a speech signal does not prove either that it
could not be done under more favorable conditions as we have shown or,
more importantly, that voiced subliminal material on tapes which "failed"
his test would not be unconsciously perceived.
While a simple spectrographic analysis of speech within a masking
sound may provide some information under the extremely favorable condi-
tions outlined above, extracting an unknown speech signal from noise is
significantly more complicated than suggested by Merikle. For an apprecia-
tion of this very difficult problem, we direct the reader to an informative
article by Gong and Haton (1787).
At the present time work from our laboratory has demonstrated a mod-
est ability to retrieve and reproduce speech-like signals from noisy envi-
ronments with SIN ratios up to -15 dB. The procedure in use is a complex
process of adaptive filtering, pitch estimation, and speech modeling. How-
ever, even this very sophisticated digital signal processing (DSP) technique

Sonograms were generated by ca turing the s eech signal ("watch out, watch out") at a sam-
;ling rate of 14.1 kHz. The totar length of & soundfile was 9 3 seconds, which was trans-
ormed from the time domain to frequency domain with an FFT (Fast Fourier Transform) in 256
k (10 msec.) segments.
522 M. J. URBAN

does not recover coherent speech. It merely provides some evidence for the
presence or absence of speech-like signals. While the performance of the fil-
tering technique in use is actually quite good, it is inadequate for the
retrieval of speech from sound tracks masked at a SIN level greater than -15
dB.

6
Frequency
(kHz)
4

6
Frequency
(kHz)
4

n
.1 .2 .3 .4 .5 .6 .7 .8 .9
Time in Seconds
FIG. 1. Sonograms: voice only (top) and masking sound only (bottom) by time (sec.)
 AUDITORY SUBLIMINAL STIMULATION: A RE-EXAMINATION 523

Egan and Hake (1950), in discussing the limits of masking sinusoidal

signals with narrowband noise, determined that such masking required a min-
imum S/N ratio of -10 dB. However, the parameters for masking a broad-
band signal such as speech with broadband noise are significantly different
from those for narrowband signals masked by narrowband noise (Krasner,

.1 .2 .3 .4 .5 .6 .7 .8 .9

Time in Seconds
FIG.1. (cont'd) Sonograms: voice and masking mix: -10 dB (top) and -15 d B (bottom)
524 M. J. URBAN

1980). Our own experience suggests that under ideal conditions, using
DSP-based methods, the present state of the art in subliminal audiotape pro-
duction when using both a broadband signal and broadband masking requires
a minimal SIN ratio between -20 and -25 dB. While this SIN ratio can be
improved upon, techniques designed to take advantage of the superior SIN

6
Frequency

6
Frequency
(kHz)
4

0
.I .2 .3 .4 .5 .6 .7 .8 .9

Time in Seconds
FIG.1. (cont'd) Sonograms: voice and masking mix: -20 dB (top) and -25 d B (bottom)
 AUDITORY SUBLIMINAL STIMULATION: A RE-EXAMINATION 525

characteristics of narrowband masking are not currently utilized in cornmer-

cia1 subliminal audiotape production. However, for research applications
- - -. we
have developed a computer-based DSP technique which does exactly that. I t
d o w s us to mask speech at levels approaching -10 dB. This represents a sig-
nificant improvement over previous methods. At the latter level of masking
speech signals are clearly available for auditory processing (Pollack, 1959)
while still remaining imperceptible or subliminal to subjects.
An additional argument against auditory subliminal processing advanced
by Merikle (1988) is that of subjective versus objective thresholds. Merikle
suggests that, although subjects claim to have no awareness of a stimulus, if
they score at better than chance on some target criterion, they must still be
aware of the stimulus. This argument is based on signal-detection theory
(Green & Swets, 1966) which implies that subjects are really aware of signals
at all levels of detectabllitv but that reporting such detection requires some
higher criteria. The somewhat tortuous logic required to allow signal-detec-
tion theory to include awareness has prompted the incisive question, "What
does it mean to claim that subjects are aware of stimuli that they deny see-
ing?" (Fowler, 1986). Recent observations on the transition between sensory
detection and sensory awareness (Libet, Pearl, Morledge, Gleason, Hosobu-
chi, & Barbaro, 1991) forces a reassessment of the signal-detection theory
paradigm with respect to subliminal stimuli, since it is clear that the process
of detection can occur separately and asynchronously with the experience of
such stimulus. Based on the compehng experimental evidence, Libet, et al.
(1991) concluded that:
When the subject reports that he feels nothing it would be a distortion of the primary evidence
to insist, on the basis of a theory, that he really felt something. The subject's introspective re-
port is the only valid primary evidence for his conscious subjective experience (p. 1752).

Libet, et al.'s findings are particularly relevant as they provide a physiologi-

cal dernonstrarion that dissociation between stimulus detection and stimulus
awareness can and does occur. Convergent evidence that this is indeed the
case comes from studies by Begleiter, Porjesz, Yerre, and Kissin (1973) and
the recent work of Taylor and McCloskey (1990). The former found that
subjective versus objective appraisals of stimuli had separate E E G localiza-
tion, and the latter demonstrated that subjects were able to perform motor
reaction-time tasks in response to a subliminal visual stimulus.
Given this persuasive evidence for unconscious processing from other
disciplines, one is led to reassess psychological explanations of unconscious
phenomena. A tentative conclusion is that the objective/subjective threshold
criteria of Cheesman and Merikle (1984), which have held sway over cogni-
tive psychology, appear to be incorrect. Also, it suggests that reliance on
verbal reports for confirmation or disconfirmation of perception is, as was
suggested by Lazarus (1956), an unreliable indicator of detection. Verbal
526 M. J. URBAN

reports seem to be reliable indicators of subjective awareness only. Although

Libet, et al.'s (1991) findings allow rejection of the Cheesman and Merikle
(1986) threshold criteria and support the notion of unconscious appraisal and
processing of sensory stimuli, it remains to be determined what complexity
of analysis occurs during this unconscious appraisal of input.
Bearing directly on this issue are the recently emerging studies in the
field of implicit memory. The implicit memory model postulates that infor-
mation which is outside the awareness of the subject and which the subject
cannot remember having acquired is used to make a response (Kihlstrom,
1987). The tacit assulnption here is that the information being utilized was
at some point in time acquired consciously. However, i t is difficult to make
that case conclusively since the subject denies prior exposure to the informa-
tion that is presently being used in some fashion. The subject is not aware
that such information is shaping the present response, so it seems difficult to
rule out the possibihty that the unconsciously utilized information was also
unconsciously acquired. A similar difficulty arises in the much more likely
case of a response being made through the simultaneous or parallel use of
u~~conscious and conscious information. How is one to determine their rela-
tive contributions? Typically in responding to the latter issue researchers have
tended to prefer explanations that defer to conscious processing by invoking
the partial-cue hypothesis (Fuhrer & Eriksen, 1960; Kellner, Butters, & Wie-
ner, 1964). In view of the more recent findings discussed above, the partial-
cue hypothesis as a general explanation for subliminal phenomena becomes
untenable.
While the research supportive of unconscious processing cited above is
not directly related to auditory subliminal stimulation, it is a parallel line of
investigation that allows us to hold in abeyance the rejection of the possibil-
ity of auditory subliminal stimulation on the basis of the criteria of Chees-
man and Merilde (1984). Although the findings of auditory psychophysics
and perception must be considered in appraising auditory subliminal stimu-
lation, they cannot be the sole criteria, either in the absence of clinical and
behavioral considerations or, more importantly, when a standardized sublimi-
nal auditory stimulus protocol is lacking.
While the work of Greenwald, et a/. (1991) and MerikJe (1988) has
played an important role in refocusing interest on the issue of auditory sub-
liminal stimulation, it has provided little evidence, in the broader context
described here, in support of the strong claim of Moore (1991) that auditory
subliminal stimulation has no effect. One might reasonably expect that such
a conclusion would have extensive evidence. In fact, this is not the case. In
contrast, the literature supporting the concept of auditory subliminal stimula-
tion as well as the efficacy of some subliminal audiotapes (much of which
goes unreported by critics) are both more extensive and more persuasive than
the two negative studies cited above.
 AUDITORY SUBLIMINAL STIMULATION: A RE-EXAMINATION 527

Because there is little distinction drawn by critics between laboratory

studies of subliminal perception and applications of auditory subliminal tech-
- -

niques as seen in the self-help industry, it is important to frame the criti-

cisms regarding self-help applications in the broader context of therapeutic
intervention. If indeed subliminal auditory self-help products must be dem-
onstrated to have therapeutic efficacy as suggested by some (Merikle, 1988),
then the proper yardstick for that evaluation would be a comparison of their
effectiveness relative to other, more conventional, interventions. Consequent-
ly, the following discussion will consider the criticisms of subliminal audio-
tape stimulation in the context of major findings in research on psychother-
apy outcome.
Therapeutic Issues
Commercial subliminal audiotapes are marketed as self-help or self-im-
provement products. They share this market niche with other self-help or
motivational types of audiotapes, videotapes, books, magazines, lectures, and
seminars. None of the latter products are required to demonstrate therapeu-
tic efficacy as their purpose is obviously not health-care delivery. Self-help is
a marketplace for functional people who are self-motivated in seeking ways
to improve their lives. There is no evidence that seriously 111 individuals
resort to self-help in lieu of professional intervention. Clinicians recognize
the enormous difference between their work as opposed to self-help strate-
gies. This is not to say that subliminal auditory stimulation as discussed here
could not be gainfully utilized in such a application, but rather that products
intended for clinical application are not generally available to consumers. '
However, since critics of commercial subliminal audiotapes (Eich, 1991;
Moore, 1991) characterize them as health-care products, it is necessary to
consider this charge and its implications.
Documenting the necessary ingredients to initiate a measurable behavior
change has been an elusive and daunting task. Since Eysenck's 1952 study
on the effects of psychotherapy, in which he found no evidence for therapeu-
tic effectiveness, psychologists have labored long and hard in their efforts
to provide evidence that was sufficiently robust to justify the application of
clinical methodologies.
The enormous difficulty in operationalizing psychotherapeutic concepts
and clinical knowledge to make them amenable to statistical constraints has
exposed a fundamental difference of philosophy between clinicians and empir-
ically oriented researchers. This clinicianlresearcher split has been well docu-

Some companies of questionable re ute will create tapes for any and all applications. These
~ r o d u c t sand titles have underminecfthe develo ment of subliminal techniques for legitimate
purposes. It is im orcanc to draw the distinct-ion getween the fact that auditory subliminal per-
cep~ionexists a n 1 the fact that subliminal audiotapes are often overpmmoted by some com-
parues.
528 M. J. URBAN

mented (Joint Commission on Mental Illness and Health, 1961) and is per-
haps best summarized by the conclusions of Bergin and Strupp (1972), that
. . . most of the standard experimental designs and statistical procedures have exerted and are
continuing to exert, a constricting effect on fruitful inquiry, and they serve to perpetuate an
unwarranted overemphasis on methodology. More accurately, the exaggerated importance ac-
corded experimental and statistical dicta cannot be blamed on the techniques proper, after aU,
they are merely tools, but their veneration mirrors a prevailing philosophy among behavioral sci-
entists which subordinates problems to methodology (p. 440).

This issue has prompted many clinicians of note, even those heavily involved
in research, to state flatly that applied research has had no effect on their
clinical practice (Bergin & Strupp, 1972). Moreover, surveys have shown that
40% of clinicians believe that no research is relevant to practice, while the
remainder believe that less than 20% of research articles have some ap-
plication to professional settings (Cohen, 1976, 1979).
I n the domain of auditory subliminal stimulation the preceding observa-
tions speak to the clinical conceptualization of the Unconscious. The Uncon-
scious has been the bane of empiricists since its introduction by Freud,
chiefly because this concept is not readily amenable to experimental investi-
gation. I t is indeed difficult, if not impossible, to construct a rigorous proto-
col to investigate the functions, processes, and limitations of the Uncon-
scious. Clinicians, for the most part, have had to come to terms with the
fact that some elements of human consciousness, motivation, and behavior
are unpredictable and perhaps unknowable. Certainly Freud (1925) suggests
that this is the case. By extension one could even argue that a theoretical
construct such as the Unconscious which, as postulated by Freud, does not
conform to the temporal or logical rules of conscious awareness cannot be
investigated by methods of conscious inquiry which are predicated as being
rational, having an external reahty and being temporally ordered. I n consid-
ering the inherent contradictions of investigating unconscious processes with
the implicit and explicit rules of consciousness, psychologists again come face
to face with the underlying and omnipresent philosophical issues of their
profession, which persist in spite of the best efforts of empiricists to quan-
tify the human experience objectively.
The ultimate resolution of this impasse may be a simple acknowledg-
ment that currently popular research strategies in cognitive psychology d o
not capture all aspects of this human experience. While the latter statement
may seem self-evident to some, it is anathema to researchers with the behav-
ioristic/empirical mind-set characterized above by Bergin and Strupp (1972).
For clinicians, primary support for the efficacy of psychotherapy was
provided by Smith, Glass, and Miller (1980) in a meta-analysis of over 500
studies, which demonstrated a treatment effect size of .63 greater than the
no-treatment group. A less well known finding of the same study was that
 AUDITORY SUBLIMINAL STIMULATION: A RE-EXAMINATION 529

placebo treatment had an effect size of .56 and so was essentially equivalent
to psychotherapy. The latter finding was reaffirmed in a study by Prioleau,
Murdock, and Brody (1983).
We mention this not as a criticism of psychotherapy, but because self-
help subliminal audiotapes are so frequently called "placebos," Indeed, given
the present state of knowledge regarding factors operative in behavioral inter-
ventions as well as profound dissatisfaction among clinicians with the empiri-
cal methods being used to assess the utility of psychotherapy (Bergin &
Strupp, 1972; Joint Commission on Mental Illness and Health, 1961), the
label of "placebo" may not even be a pejorative criticism. In a preliminary
report (Costello & Budzynski, 1991) on the utilization of commercially pro-
duced subliminal audiotapes with a clinical population, the tape-users' scores
of improvement on several outcome measures were significantly better than
those of the control group and were equivalent to those of the psychotherapy
group.
Swingle (1992) documents ten years of work with subliminal auditory
stimulation in a wide variety of successful experimental protocols. Currently
he uses subliminal audiotapes as adjuncts in clinical treatment. Similar work,
supported by the compehng physiological evidence of EEG correlates to sub-
liminal stimulation, has been conducted and documented by several investi-
gators (Dixon & Lear, 1963; Emrich & Heinemann, 1966; Shevrin & Dick-
man, 1980).
In defense of subliminal techniques as possible adjuncts in behavioral
interventions we offer the following prescient statement which was made by
Hyman and Berger in response to Eysenck's (1966) criticism of psychother-
apy:
Let us not, under the impulse of urgency and a misguided sense of the "scientific" pretend we
can settle the issue today if only we apply the correct M E T H O D O L O G Y and make crude anal-
ogies between successful procedures in science and field studies in therapy (p. 86).

Although Eich and Hyman (1991), when assessing subliminal audio-

tapes, have seemingly forgotten Hyman's admonition, it appears that this
statement applies as much to the emerging technology of today's procedures
for auditory subliminal stimulation as it did to psychotherapy in 1966.
Since subliminal input is outside the awareness of the receiver, the
effect of such messages is even more difficult to assess than the previously
described difficulties associated with the conscious interventions of psycho-
therapy. As Poetzl described in 1917 (19601, subliminal stimulation is often
manifest in the dreams of subjects exposed to such stimulation. These mani-
festations sometimes took several weeks to appear, so we are confronted, as
researchers, with the formidable task of attempting to define the correspon-
dence of a subliminal stimulus input to some target output of cognitive, af-
fective, or behavioral dimension that has been processed through unknown
strata of mental processes for some indeterminant period of time.
530 M. J . URBAN

This is a situation similar to what investigators face in psychotherapy

outcome studies. The point here is to highlight the inherent contradiction of
requiring subliminal perception, which most researchers regard as a restricted
information channel, to demonstrate a level of efficacy that is not evidenced
by any of the currently popular and widely used methods of psychotherapy.
Logic as well as a grasp of the literature on behavior change dictate that we
either afford subliminal stimulation the same latitude granted psychotherapy
(Hyman & Berger, 1966; Andrews & Harvey, 1981; Landman & Dawes,
1982) or we unilaterally adopt Eysenck's position that all psychotherapy is
ineffective. Adoption of the latter point of view is unlikely in view of the
acknowledged pragmatic utility of psychotherapeutic intervention. The con-
sensual viewpoint is that the difficulty with demonstrating the usefulness of
psychotherapy lies in the complexity of the entire gestalt of the therapeutic
process rather than with the lack of therapeutic effect.
Based on a detailed overview of subliminal stimulation, a similar, albeit
amplified, situation occurs for the demonstration of subliminal utility. This
difficulty can lead, as perhaps in the Greenwald, et al. (1991) study, to a
Type I1 error (incorrectly retaining the null hypothesis) or a tendency to mis-
interpret findings according to preconceived theoretical bias, called by Melt-
zoff and Kormeich (1970) the "alibias."
I n summary, it appears that hasty dismissal of subliminal auditory stim-
ulation is both premature and dl-advised, particularly in consideration of the
potential clinical benefits to be derived from the availability of an effective
auditory subliminal stimulus delivered by way of a standardized audiotape
presentation either as therapeutic tools or as therapeutic adjuncts.
We now turn to a brief consideration of technical issues of critical im-
portance in the production of subliminal audiotapes, an area which is typi-
cally overlooked in critical appraisals. Without an understanding of technical
parameters, assessing the efficacy of subliminal audiotapes is not possible.
This is an issue of some importance, as the lack of standardization evident in
work on subliminal psychodynamic activation (Fudin & Benjamin, 1991) also
prevails throughout the self-help industry. This has led to the availability of
products that vary widely in quahty of production as defined by SIN ratios.
A global assessment of commercially available subliminal auditory tape prod-
ucts is not possible.
Technical Issues
Notably absent in the publications of Greenwald, et al. (1991) and of
Merikle (1988) is any consideration of technical issues attendant to the pro-
duction of audiotaped subliminal products or the rather obvious fact that
commercial products are decidedly not generic irrespective of their often-
times similar advertising claims. As has previously been reported by Zen-
hausern, et al. (1973), Zenhausern and Hansen (1974), and more recently
 AUDITORY SUBLIMINAL STIMULATION: A RE-EXAMINATION 53 1

elaborated by Swingle (1992), certain minimal SIN ratios must be achieved

to optimize perception by this technique.' The early subliminal processors
popularized by Hal Becker and known as the "black box" used a simple cir-
cuit in which the level of the masking signal attenuated the gain of the
message, whereby maintenance of a fairly constant decibel difference be-
tween the message and the maslung sound was attempted. Given the param-
eters for broad-band masking cited by Krasner (1980), this device probably
had an optimum performance in the neighborhood of -30 dB. However, the
attenuation could be set by the user. In one study that used this system in a
double-blind design, the sound engineer indicated that the average S/N ratio
was approximately -50 dB. The authors never mentioned this high S/N ratio
as possibly contributing to their null results (Lenz, 1989). For reference pur-
poses, -50 dB is approaching the noise floor of magnetic tapes, the sound
that is referred to as tape "hiss." A second oft-cited study on subliminal
audiotapes which had a negative outcome was conducted by Russell, Rowe,
and Smouse (1991). Nor reported by the authors, who may have been
unaware of the fact, is that the tapes used in this study were not produced
with a subliminal mixer of the Becker-type described above. Instead, the
tapes were produced in a sound studio by simply doing a multitrack record-
ing and attenuating the message channel track to a level at which it became
inaudible. Since this method lacks the amplitude-following features of the
Becker unit, it is necessary to attenuate the message much further to achieve
complete masking. Since the voice is not following the masking sound, the
SIN ratio becomes variable, so the masking level conceivably could vary be-
tween -30 and -60 dB. While these issues do not invalidate the findings of
the Russell, et al. (1991) study, ir would be speculative to suggest that their
findings for this method of subliminal stimulus delivery generalize to aU
audiotape methods of subliminal activation.
While there may be some message retrieval by the brain under these
conditions, present evidence suggests that SIN ratios greater than -30 dB
show a marked degradation in effectiveness. The Becker unit also processed
signals too slowly to ensure masking at aU times. Due to limitations of the
hardware, there were often "overshoot" conditions in moments when the
masking sound became quieter, such as with ebbing surf or musical rest pe-
riods. At these times the circuitry could not compensate with sufficient
speed to prevent the audible appearance of sibilants or fragments of words.
As mentioned previously, this was typically compensated for by increasing
the masking ratio to as much as -50 dB.

'While much of Swingle's (1992) work with ararnetric definitions refers to S/N ratios as a
function of difference in sound-pressure level getween source and subject, his corollary work
with the more typical magnetic tape suggests that the two formats can achieve virtually equiva-
lent results.
532 M. J. URBAN

Another vital and often ignored issue is that of tape content. The
scripts or affirmations which comprise the instructional messages of these
products are key ingredients of their effectiveness. By way of example, con-
sider the following hypothetical affirmations:
(1) I never think badly of myself.
(2) I think well of myself.

For purposes of our argument we will consider these two statements to

be essentially equivalent when assessed by the conscious mind. However,
when presented as subliminals, they are quite different and indeed, sentence
one, when presented subliminally, could conceivably induce a negative re-
sponse in spite of its positive intent. While space constraints do not allow
for a full consideration of scripting as an element of the differential success
or utility of various commercial products, it should be noted that commercial
products are not homogeneous with respect to scripts. I n brief, good script-
ing using short phrases or words of emotionally relevant material is more
likely to be influential than neutral words or phrases (Dixon & Henley,
1991; Shevrin, 1980; Swingle, 1992). However, because many questions
must yet be answered with respect to level of analysis that is provided by
unconscious processes, there is no absolute criterion for what a reliable stim-
ulus word would be. Based on some studies in perceptual defense (Kurland,
1954; Groeger, 1986, 1988), one would infer that individual responses to
auditorily presented subliminal words would vary with each subject.
Most subliminal audiotapes utilize some method of 'hiding' the speech
signal. There are many techniques employed in accomplishing this since the
term 'subliminal' as it applies to auditory stimulation has come to be used as
an imprecise descriptive term meaning 'not available to conscious perception.'
The most common ways in which to mask or disguise speech include
masking by broad-band noise, music, or psychoacoustic concealment. More
recently, we have also seen the evolution of the so-called "silent subliminal"
which is no more than rhe voice message frequency-modulated with a carrier
near the upper end of the range of human hearing, typically around 15 kHz.
This is an interesting concept which requires that the brain function as a
demodulator. This may or may not be a part of the brain's repertoire, but at
the present time evidence to support such a contention is rather sparse. With
respect to airborne conduction of ultrasound, a single intriguing study by
Martin, Hawryluk, and Guse (1974) described effects of ultrasonic tones on
behavior in several experimental protocols. While pure tones are quite differ-
ent from speech that has been modulated into the ultrasonic range, the
observation that an undetectable ultrasonic stimulus can have a behavioral
effect remains interesting. With respect to the larger question of demodulat-
ing ultrasonic speech, data from a study on bone conduction of sound shows
 AUDITORY SUBLIMINAL STIMULATION: A RE-EXAMINATION 533

that the brain is able to decode speech when these signals are modulated into
the ultrasonic (>24 kHz) frequency range (Lenhardt, Skellet, Wang, &
Clarke, 1991). Although the speech signals in the latter study are not air-
borne, the mere fact that the brain is able, through some unidentified
neuronal substrate, to carry out the complex temporal and spectral discrimi-
nation required for speech perception under these conditions, is in itself very
provocative. Whether some method could demonstrate similar effects via air
conduction is presently under investigation. I n unpublished work from this
laboratory, we have conducted preliminary investigation into the possibility
that speech presented at the upper end of the audible spectrum may be avail-
able as another means of presenting an auditory stimulus outside of con-
scious perception. This has been accomplished by both frequency and ampli-
tude modulation of speech s~gnalsat frequencies in excess of 14 kHz and
then recording them onto magnetic tape. Using GSR and H R measures to
monitor the effects of such a modulated signal (original voice message was:
"Watch out, watch out, danger!," spoken in a panicked voice), we observed
changes in both measures coincident with the stimulus presentation. While
we draw no conclusions from these observations, it suggests another possible
avenue of investigation for developing an auditory 'subliminal' stimulus.
Each of the methods described above has its own constraints; any
broad-band maslung sound such as white noise, pink noise, or ocean surf
must employ the minimal possible maslung ratio between the noise and the
signal to facilitate message retrieval. Music as a masking medium has the spe-
cial problem of rest moments in which there are no sounds and by extension,
no message. Here the problem is one in which the music is cutting off the
message at an unknown point, thereby perhaps modifying the intent of the
original statement. We d o not yet know whether psychoacoustic conceal-
ment, in w h c h the voice is made to sound like part of the dominant sound
track or like a particular instrument, is an effective method of stimulus de-
livery.
I n a studio setting each of the methods outlined above can be further
enhanced by multitracking, track "bouncing," compression, or speeding. It
is through the use of such techniques that one can develop a commercial tape
which boasts a million affirmations. Claims of this nature have raised crit-
icisms from some engineers. These criticisms had to d o with the band-width
restrictions that arise with pitch increases consequent to tape speeding, since
there is a linear relationship between pitch and bandwidth. I n analog record-
ing this means that a point exists at which a speeded message will utllize
the entire available bandwidth of an audiotape. Nominal parameters suggest
that bandwidth restrictions using analog equipment would limit speeding to
a maximum factor of six. However, pitch-shifting and the resulting band-
334 M. J. URBAN

width restrictions can be minimized if recording is done at a sampling rate of

44.1 kHz in the digital domain.
Moreover, since speeding is not necessarily the only way to place large
numbers of affirmations on magnetic media, this articular argument against
the "million affirmation" tape is incorrect. Another technique for increasing
the number of affirmations on a tape is to compress the speech, a process in
which parts of words and pauses are removed. This too is easily done with
digital signal-processing methods and does not result in pitch shifting. How-
ever, there are other problems associated with placing this many affirma-
tions on a magnetic medium that are even more basic and which render the
putative effectiveness of such a tape rather doubtful. Chief among these is
the problem of r e t r i e ~ a b i l i t ~ .
The notion that "if a little is good, then a lot is better" may not be
true with respect to subliminal audiotapes. Given that some investigators sug-
gest that subliminal stimulation yields a weak effect in a low-capacity system
(Dixon & Henley, 1991), logic would seem to favor minimal tracks with
maximal clarity. This is, in fact, what Swingle has demonstrated (1992). So,
while it is possible to place an almost infinite number of messages on a tape,
there is little evidence that this is superior to a few tracks minimally masked
and produced with high fidelity. The really critical questions are not how
much can be put on a tape, but what can be retrieved, both within the con-
straints of playback equipment and human perceptual abilities, and most
importantly how will the unconscious perception, appraisal, and use of such
information be manifest, either as a component of a subject's private internal
dialogue or as some externaly measurable quality. The latter questions can
only be properly answered if the technical issues are attended to prior to
embarking on such investigations.
With respect to speeding, it is important to realize that this can now be
done digitally on a computer, which, as mentioned above, easily overcomes
the problem of pitch shifting that occurs with analog equipment. I t has often
been erroneously reported (Moore, 1991) that a speeded speech track sounds
like a high-pitched squeal. I n the absence of qualifying conditions (e.g.,
'when using analog equipment'), such statements reflect an outdated percep-
tion of recording techniques and inaccurately portray the present state of
recording technology.
I n our present work we have overcome all of these problems by using a
computer-based commercial DSP board, which has the processing power and
speed necessary to prevent the occurrences just mentioned. Also, we have
developed software algorithms that maintain a constant S/N ratio by simulta-
neously 'looking at' the instantaneous amplitude of both the signal and the
masking sound, thereby reliably maintaining a constant S/N ratio. Further,
the speed of the DSP chip allows us to use a time window as small as
 AUDITORY SUBLIMINAL STIMULATION: A RE-EXAMINATION 535

.00002 sec., which is several orders of magnitude better than the original
"black box," and in fact allows the digital subliminal processing and filtering
to be accomplished in real time. To minimize the S/N ratio further it is also
possible a n d often helpful to filter out some of the high frequency speech
components prior to the subliminal mixing. A useful approach is to perform
initial spectral analyses of both the signal and noise, which allows selection
of filters to optimize the effectiveness of the masking sound. It is, after all,
essential that the frequency components present in the message are also
present in the masking medium. Based on the considerations described
above, one can fairly say that the correctly produced present-day audiotape
subliminal stimulus bears little relationship to its predecessors of 20 years
ago other than in concept and in name. This crucial fact is both unrecog-
nized and unacknowledged by present-day critics of subliminal audiotapes.
More importantly these recent innovations give researchers the opportunity
to begin investigating this phenomenon with standardized subliminal stimu-
lus formats. This should lead to the determination of what sorts of research
protocols yield the elusive effects of auditory subliminal stimulation.
Let us now turn our attention to a brief review of current investigations
that have utilized subliminal audiotape techniques, in some cases commer-
cially available tapes, in their research protocols.
Current Research Results
I t is generally conceded that visual subliminal perception is a valid phe-
nomenon (Hardaway, 1990; Bornstein, 1990; Greenwald, Klinger, & Liu,
1989). However, opponents of audiotape subliminal stimulation tend to ig-
nore the important Literature in cross-modal subliminal processing (Baker,
1936; Henley, 1975; Hardy & Legge, 1968), which suggests that both sen-
sory modalities share equivalent capabilities with respect to subliminal pro-
cessing, and, more importantly, are interactive in that domain. This, howev-
er, is a theoretical point of issue and is not to be discussed further, although
it does have some interesting implications for future work.
I n the domain of auditory subliminal perception, a significant body of
information is frequently neglected. One of the more significant omissions
concerns the book, Subliminal Treatment Procedures: A Clinicians' Guide, by
Swingle (1992). Swingle documents a decade's worth of experimentation in
which he defines the benchmark parameters of auditory subliminal stimula-
tion as well as providing substantial experimental and clinical evidence that
these procedures can afJect emotions, problem solving and memory, aesthetic
judgement, task performance,
- . general arousal, and interpersonal behavior. From
a clinical perspective, these findings suggest some exciting possibilities.
I n addition to Swingle's work, there is now a great deal of converging
evidence for the utility of auditory and visual subliminal stimulation, which
has led Dixon and Henley (1991) to suggest that clinicians begin using the
536 M. J. URBAN

information derived from academic research in clinical practice.' Dixon and

Henley summarize the current state of affairs in subliminal stimulation
research as follows:
From the results of applying subliminal stimulation to problems of diagnosis and therapy, three
conclusions may be drawn. First, the data support the view that the meaning of external stimuli
of which the recipient is unaware may be responded to and determine emotional responses, lexi-
cal decisions, overt behavior, and subjective experience. Second, they confirm the reahty of
psychopathology, that is, of a substrate of emotionally colored stored information with a poten-
tial for producing somatic symptoms and disorders of thinking, affect, and behavior. Third, to
the extent that the content of psychopathology is screened From conscious scrutiny and therefore
impervious to supraliminal information, so may it be accessed and ameliorated by drive-related
stimuli of which the subject is not aware (p. 250).

Although the combined reports of Swingle and Dixon and Henley neu-
t r d z e many of the standard objections lo subliminal auditory stimulation,
also available as confirming evidence for auditory subliminal effects are those
studies in which tapes similar to the commercial products available to con-
sumers have been used.
A recent study described the preliminary results of psychotherapeutic
intervention by subliminal audiotape alone (Costello, et al., 1991). In this
preliminary study the authors tracked the progress of 63 psychotherapy cli-
ents over an average time of eight months at the Cassel Research Center in
Mornington, Australia. Each of these patients was treated exclusively with
subliminal audiotapes of their choosing. Their progress was tracked through
use of the Lifestyle Analysis Test, a previously vahdated instrument (Gilley
& Uhlig, 1985). This preliminary report examined only the changes on the
self-esteem subscale of the test instrument, comparing improvement gain dif-
ferences in the pre- and posttest scores between subliminal tape therapy vs
control group subjects. Repeated-measures analysis of variance resulted in an
omnibus F,,8, of 18.14 ( p < .0001). Post hoc analysis utilizing the very con-
servative and robust Games-Howell procedure (1976) indicated a significant
difference between the subliminal tape users versus the no-therapy control
group on the pre- and posttest measures ( p < .01).
Since previous researchers (Cassell & Costello, 1991) have shown the
self-esteem subscale of the Lifestyle Analysis Test to be positively correlated
with social involvement, assertiveness, and satisfaction, a combination which
is a global measure of Ego Strength, the present preliminary report tentative-
ly suggests that subliminal audiotape intervention may be a positive treat-

'Questions have been raised regarding ethical implications of auditory s u b h l n a l products, with
suggestions that they be regulated. This concern, while legitimate, seems premature in light of
the still existing controversy over whether auditory subliminal stimulat~on1s a real henomenon.
Pendmg resolution of this issue, a possible safeguard would be for clinicians to gecome more
directly involved in the self-help industry to establish quality control so available products are
both ethical and consistent with current thinking in behavioral interventions.
 AUDITORY SUBLIMINAL STIMULATION: A RE-EXAMINATION 537

ment modality, either alone or adjunctively, as deemed appropriate by the

attending professional, in the remediation of some psychological problems.
In another pilot study (Pellca, Taylor, & Fedrigotti, 1991), subliminal
audiotapes were utilized in a weight loss protocol. The 34 subjects in this
study experienced an over-all weight loss of 6.4 kg in a nine-week period. In
spite of these encouraging results, the methodology for the pilot study was
quasi-experimental and so the results await confirmation from the large-scale,
double-blind follow-up which is currently underway.
I n a study which used subliminal audiotapes to alter perceptual order
formation in vision, Kruse, Stadler, and Kobs (1992) successfully biased sub-
jects' perception of apparent motion through the presentation of subliminal
auditory cues. The tapes used in this study were produced with a subliminal
mixer similar to what was described above. The experiment used ambiguous
apparent-motion displays (Kruse, 1988) which can be perceived in three al-
- -

ternative ways. The addition of subliminal suggestions introduced a selection

bias in the subjects' perception of the multistable display.
In another application of interest, Monahan (1991) demonstrated the
usefulness of subliminal audiotapes (again produced with a subliminal mixer,
SIN<-30 dB) as an adjunct in a 12-step treatment program for chemical
dependency. Using the Profile of Mood States as an outcome measure in a
double-blind design, subliminal tape users' improvement on the global out-
come measure, the Total Mood Distortion Scale, was better than both the
placebo and the control groups' scores.
Reid (1990) used subliminal audiocassettes as an adjunct in the treat-
ment of depression. His findings are interesting in that they showed a dosage
relevancy in exposure time. Patients using the subliminal audiotapes did not
show any improvement gains over the control group until they had been
exposed to the tapes for more than 15 hours. At that point an analysis of
improvement as a-function of exposure to the subliminal tape showed a sig-
nificant change on Beck Inventory scores (p = .002), improved daily mood
scores (p = .04), and enhanced the therapists' rating of therapy ( p = .009). All
of these findings were observed in comparisons with a control group who
also were in therapy but who were listening to a placebo tape under dou-
ble-blind conditions.
Borgeat, Boissonneault, Chaloult, and Elie (1989) partially replicated
the findings of their earlier report (Borgeat & Goudet, 1983) in which audi-
tory subliminal stimuli were found to activate physiolog~calresponses as mea-
sured by GSR and heart rate. I n the current study activation was only found
on the heart-rate parameter. There is, however, some question as to whether
the verbal stimulus in this study was completely subliminal. As described,
the authors indicated that the activating suggestions were presented at 25 dB
and masked by a 40-dB white noise. This implies a S/N ratio of -15 dB,
538 M. J. URBAN

which appears too low to ensure complete masking of speech (see discussion
above), hence leaving open the possibility of partial cueing. I n contrast to
the preceding study, Kotz6 and Moller (1990) reported a significant increase
in GSR response to emotional words presented as an auditory subliminal
stimulus. Here again, the stimulus was produced with a subliminal mixer,
with the voice message some 3 to 5 dB below threshold, which suggests a
SIN ratio between -25 to -30 dB.
In their totality these studies represent support both for subliminal audi-
tory perception as a physiological phenomenon and for the presentation of a
subliminal stimulus via the audiotape medium. Consideration of the above
cited studies in conjunction with the existing body of literature in subliminal
perception plus the substantial amount of converging evidence for auditory
subliminal perception from interdisciplinary sources suggests that continued
efforts to reject the auditory modality while accepting the visual modality as
a legitimate pathway for subliminal perception are unwarranted, require tor-
tuous logic, and lead to physiological inconsistencies.
To summarize, the following minimal issues should be addressed when
investigating auditory subliminal stimulation: a standardized method of stim-
ulus presentation, a reliable method of masking subliminal messages, and the
maintenance of a minimal signal-to-noise ratio in stimulus presentation. Per-
haps this overview w d help revive some much-needed research interest in
auditory subliminal phenomena among investigators in both the academic
and clinical community.
REFERENCES
ANDREWS,G., & HARVEY, R. (1981) Does psychotherapy benefit neurotic patients? A reanaly-
sis of the Smith, Glass and Miller data. Archives of General Psychiatry, 38, 1203-1208.
BAKER,L. (1937) The influence of subliminal stimuli upon verbal behavior. ]ournu/ of Ewperi-
mental Prychology, 20, 84-100.
BECKER,H. C., CORRIGAN, R. E., ELDER,S. T., TALL ANT,^., & GOLDSTEIN,M. (1965) Sublim-
inal communication: biological engineering consi eratlons. Proceedings of the 6th In-
ternational Conference of Medical Electronics and Biological Engineering, Tokyo.
BEGLEITER,H., PORJESZ,B . , YERRE,C., & KISSIN,B. (1973) Evoked potential correlates of ex-
pected stimulus intensity. Science, 179, 814-818.
BEN-HUR,A. (1979) The relationshi of systematic desensitization and the activation of s mbi
otic merging fantasy to speec{ anxiety reduction amon college students. ( ~ n ~ u b l s h e d
doctoral dissertation, New York Univer.) Dissertation
2352B.
ist tracts
International, 40, 2351B-

BERGIN,A . E., & STRUPP,H. H. (1772) Changing frontiers in the science of psychoiherapy. New
York: Aldine.
BORGEAT, F., BODSONNEAULT, J., CHALOULT,L., & ELIE, R. (1989) Psychophysiolo ical re-
sponses to subliminal auditory suggestions for activation. Perceptual and Motor ~fills,69,
947-953.
BORGEAT,F., & GOULET, J. (1983) Psychophysiological changes following auditory subliminal
suggestions for activation and deactivation. Perceptual and Mofor Skills, 56, 759-765.
BORNSTEIN,R. F. (1990) Critical importance of stimulus unawareness for the production of sub-

6This author will make subliminal audiotape products conforming to the specifications described
in this paper available to qualified investigators.
 AUDITORY SUBLIMINAL STIMULATION: A RE-EXAMINATION 539
liminal psychodynamic activation effects: a meta-analytic review. Journal of Clinical Psy-
chology, 46, 201-210.
BOUCHARD,S. J. (1984) Effects of self-administered subliminal-relaxation treatment on anxiety.
Dissertation Abstracts International, 45, 1096B. (University Microfilms No. DA8420101)
BUCHHOLZ, E. S. (1969) A study in the management of aggression by schizophrenics: the ef-
fects of aggressive stimuli on an aspect of cognitive functioning in schizophrenics and
normals. (Unpublished doctoral dissertation, New York Univer.) Dissertation Abstracts In-
ternational, 30, 376B-377B.
CASSELL,R. N., & COSTELLO,B. R. (1991) Developing a life style syntalics for persons in-
volved in psychological therapy. College Student]ournal, 25, 215-221.
CHEESMAN, J., & MERMLE,P. M. (1984) Priming with and without awareness. Perception and
Psychopbysics, 36, 387-395.
CHIMERA,D. M. (1987) An exploration of the effect of auditory subliminal stimuli on schizo-
enic pathology. Dissertation Abstracts International, 48, 1509B. (University Micro-
f'+
llms No. DA 8718326)
COHEN,L. H . (1976) Clinicians' utilization of research findings. JSAS Catalog of Selected Doc-
uments in Psychology, 6 , 116.
COHEN,L. H. (1979) The research readership and information source reliance of clinical psy-
chologists. Professional Psychology, 10, 780-786.
COS~ELLO, B. R., & BUDZYNSKI, T. H . (1991) Subliminal audiotapes as a psychotherapeutic mo.
dality. Paper presented at the annual meeting of the International Society for Precon-
scious Learning, Las Vegas, Nevada. (manuscript in preparation)
DIXON,N. F. (1981) Preconscious processing. New York: Wiey.
DIXON,N. F., & HENLEY,S. H. A. (1991) Unconscious perception: possible implications of
data from academic research for chical practice. Journal of Nervous and Men&[ Dimse,
179, 243-252.
DIXON,N. F., & LEU, T. (1963) Electroencephalograph correlates of threshold regulation.
Nature, 198, 870-872.
DOCHE-BUDZYNSKI, L., & BUDZYNSKI,T. (1989) Subliminal self-esteem enhancement in adult
Type A males. Education, 10, 50-55.
EGAN,J., & HAKE,H . (1950) On the masking pattern of simple auditory stimulus. Journal of
the Acoustics Society of America, 22, 622-630.
EICH, E. (1991) Subliminal pseudoscience. Paper resented at the annual convention of the
Committee for the Scientific Investigation o/claims of the Paranormal, Berkeley, CA.
EICH,E., & HYMAN, R. (1991) Subliminal self help. In D. Druckrnan & R. Bjork (Eds.), In
the mind's eye: enhancing human performance. Washington, DC: National Academy Press.
Pp. 107-119.
EMRICH,H., & HEINEMANN, L. G. (1966) EEG bei unterschwelliger Wahrnehmung emotional
bedeutsamer Worter. Psychologische Forschung, 29, 285-296.
ERIKSEN,C. W. (1956) Subception: fact or artifact? Psychological Review, 63, 74-80.
ERIKSEN,C. W. (1960) Discrimination and learning without awareness: a methodological sur-
vey and evaluation. Psychological Rewiew, 67, 279-300.
EYSENCK, H. J. (1952) The effects of psychotherapy: an evaluation. Iournal of Consulting and
Clinical Psychology, 16, 3 19-324.
EYSENCK, H. J. (1966) The effects ofpsychotherapy. New York: International Science Press.
FOWLER,C. A. (1986) An operational definition of conscious awareness must be responsible to
subjective experience. Behavioral and Brain Sciences, 9, 33-35.
FREUD,S. (1925) The Unconscious. In The Standard Edition of the Collected Papers of Sigmund
Freud. [ J . Rivlere (Ed.)] Vol. 4. London: Hogarth Press. Pp. 98-136.
FUDIN,R., & BENJAMIN,C. (1991) Review of auditory subliminal psychodynamic activation ex-
periments. Perceptual and Motor Skills, 73, 1115-1136.
FUHRER, M., & ERIKSEN,C. W. (1960) The unconscious perception of the meaning of verbal
stimuli. Journal of Abnormal and Social Psychology, 61, 432-439.
GAMES,P. A,, & HOWELL,J. F. (1976) Pairwise multiple-corn arison procedures with unequal
n's and/or variances: a Monte Carlo study. Journal of ~ B u c a t i o n aStatistics,
~ 1, 113-125.
GILLEY,W. F,,& UHLIG, G. E. (1985) Validation of Cassel Type-A Personality Assessment
Profile. Psychology, 22, 4- 10.
540 M. J. URBAN

GONG,Y., & HATON,J-P. (1987) Time domain harmonic matching pitch estimation using time-
dependent speech modeling IEEE Transactions on Acoustics, Speech and Signal Process-
ing, ASSP-35, 1386-1400.
GREEN,D. M., & SWETS,J. A. (1966) Signal detection theory and psychophysics. New York:
Wilw .
GREENWALD,A. J., KLINGER,
M. R., & LIU, T. J. (1989) Unconscious processing of dichopti-
cally masked words. Memory and Cognition, 17, 35-47.
GREENWALD, E. R., PRATKANIS,
A. J., SPANGENBERG, J. (1991) Dou-
A . R., & ESKANAZI,
ble-blind tests of subliminal self-help audiotapes. Psychological Science, 2, 119-122.
J. A. (1986) Predominant and non-predominant analysis: effects of level of presenta-
GROEGER,
tion. British Journal of Psychology, 77, 109-116.
GROEGER,
J. A. (1988) Qualitatively different effects of undetected and unidentified auditory
primes. Journal of Experimental Psychology, 40, 323-329.
R. A. (1990) Subliminally activated symbiotic fantasies: fact and artifacts. Psy-
HARDAWAY,
chological Bulletin, 107, 177-195.
HARDY,G. R., & LEGGE, D. (1968) Cross-modal induction of changes in sensory thresholds.
Quarterly Journal of Ewperimental Psychology, 20, 20-29.
HENLEY,S. H . A. (1975) Cross-modal effects of subliminal verbal stimuli. Scandinavian Journal
of Psychology, 16, 30-36.
HENLEY,S. H. A. (1984) Unconscious perception re-visited: a comment on Merikle's (1982)
paper. Bulletin ofthe Psychodynamic Society, 22, 121-124.
HOLENDER, D. (1986) Semantic activation without conscious identification in dichotic listen-
ing, parafoveal vision, and visual masking: a survey and appraisal. Behavioral and Brain
Sciences, 9, 1-66.
HYMAN,R., & BERGER,L. (1966) Discussions. In H. J. Eysenck (Ed.), The effects of psycho-
therapy. New York: International Science Press. Pp. 81-86.
JOINTCOMMISSION ON MENTALILLNESSAND HEALTH.(1961) Action for mental health. New
York: Science Editions.
K A ~ E RV., A. (1986) The effects of an auditory subliminal message upon the production OF
images in dreams. Journal of Nervous and Mental Disease, 174, 397-407.
KELLNER, H., B ~ R s N., , & WIENER, M. (1964) Mechanisms of defense: an alternative re-
sponse. Journal of Personality, 32, 601-622.
KIHLSTROM, J. F. (1987) The cognitive unconscious. Science, 237, 1445-1452.
K o ~ z t ,H. F., & MOLLER,A. T. (1990) Effect of auditory subliminal stimulation on GSR.
Psychological Reports, 67, 931-934.
KRASNER, M. (1980) The critical band coder: digital encoding of speech signals based on the
perceptual requirements of the auditory system. Proceedings of IEEE International
Conference on Acoustics, Speech, and Signal Processing, Denver, CO. Pp. 327-331.
KRUSE, P. (1988) Stabilitat, Instabilitat, Multistabilitat: Selbstorganisation und Selbstrefer-
entialitat in kognitiven Systemen. Delfin, 11, 35-57.
KRUSE,F!, STRADL~R, M., & KOBS,M. (1992) Suggestion and perceptual instability: auditory
subliminal influences. Journal o/ Cognitive Psychology, in press.
KURLAND, S. ( l 9 j 4 ) The lack of generality in defense mechanisms as indicated i n auditory per-
ception. Journal of Abnormal and Social Psychology, 49, 173-177.
LANDMAN, J. T., & DAWES,R. M. (1982) Psychotherapy outcome: Smith and Glass' conclu-
sions stand up under scrutiny. American Psychologirt, 37, 504-516.
LAZARUS, R. S. (1956) Subception: fact or artifact? A reply to Eriksen. Psychological Review,
63, 343-347.
LAZARUS, R. S., & MCCLEARY, R. (1951) Autonomic discrimination without awareness: a study
of subception. Prycbological Review, 58, 113-122.
LENHARDT, M., SKELLET, R., WANG,P., & CIARKE,A. (1991) Human ultrasonic speech per-
ception. Science, 253, 82-84.
LENZ,S. (1989) The effects of subliminal auditory stimuli on academic learning and motor
skills performance among pol~cerecruits. Unpublished doctoral dissertation, California
School of Professional Psycholog)
LIBET,B., PEARL,D. K., MORLEDGE, D E , GLEASON, C. A., HOSOBUCHI, Y., & BARBARO, N.
 AUDITORY SUBLIMINAL STIMULATION: A RE-EXAMINATION 541
M. (1991) Control of the transition from sensory detection to sensory awareness in
man by the duration of a thalamic stimulus. Brain, 114, 1731-1757.
MARTIN,D. G., HAWRYLUK, G. A,, & GUSE,L. L. (1974) Experimental study of unconscious
influences: untrasound as a stimulus. Journal of Abiiormal Psychology, 83, 589-608.
MELTZOFF, J., & KORNREICH, M. (1970) Research in psychotherapy. New York: Atherton.
MERMLE,P. M. (1988) Subliminal auditory tapes: an evaluation. Psychology 6 Marketing, 46,
355-372.
MODESTIN, J. (1985) Psychological components of psychopharmacological side effects. Bulletin
of the Menninger Clinic, 49, 29-36.
MOERTEL,C. G., TAYLOR, W. F., ROTH, A,, & TYCE,F. (1976) Who responds to sugar pills?
Mayo Clinic Proceedings, 51, 96-100.
MONAHAN, R. (1991) Subliminal audio as an adjunct to traditional twelve step oriented treat-
ment for chemical d e p e n d e n ~ ~ - ~ i l oproject.
t Livegrin Foundation, Philadelphia, PA.
MOORE,T. E. (1991) Subliminal auditory self-help tapes. Paper presented at the APA Con-
vention, San Francisco, CA.
PELKA,R., TAYLOR, E., & FEDRIGOT~I, T. (1992) Application of subliminal therapy to over-
weight subjects. Behavioral Medicine, in press.
POETZL,0 . (1917/1960) The relationship between experimentau induced dream images and
indirect vision. Psychological Issues, 2, 4 1-120. (Original pugl. 1917)
POLLACK,I. (1959) Message uncertainty and message reception. Journal of the Acoustical
Society of America, 31, 1500-1508.
PFZOLEAU, L., MURDOCK, M., & BRODY,N. (1983) An analysis of psychotherapy versus placebo
studies. Behavioral and Brain Sciences, 6, 275-310.
REID, J. (1990) Free of depression: subliminal tape study. Unpublished master's thesis, Col-
orado State Univer.
RUSSELL,T., ROWE, W., & SMOUSE,A. (1991) Subliminal selE-help tapes and academic
achievement: an evaluation. Journal of Counseling and Development, 69, 359-362.
SHEVRIN,H., & DICKMAN, S. (1980) The psychological unconscious: a necessary assumption
for all psychological theory? American Psychologist, 35, 421-434.
SHIFREN,I. E. W. (1981) The interaction between hemispheric preference and the perception
of subliminal auditory and visual symbiotic gratification stimuli. Dissertation Abstracts
International, 42, 4211B-4212B. (University Microfilms No. DA 8202698)
SILVERMAN, L. H. (1978) Unconscious symbiotic fantasy: a ubiquitous therapeutic agent.
International Iotrrnal of Psychoanalytic Psychotherapy, 7 , 562-585.
SMITH, M. L., GLASS,G . V., & MLLLER,T. I. (1980) The benefits of psychothwapy. Baltimore,
MD: Tohns H o ~ k i n sUniver. Press.
STROFI,M.,SHAM,A . , & WASHBURN, M. (1908) A study in guessing. American Journal of
Psychology, 19, 243-245.
SWINGLE,P. G. (1992) Subliminal treatment procedures: a clinicians' guide. Sarasota, FL: Pro-
fessional Resource Press.
TAYLOR, J. L., & MCCLOSKEY, D. I. (1990) Triggering of preprogrammed movements as reac-
tions to masked stimuli. Journal of Neurophysiology, 63, 439-446.
Vance us Judas Priest. (1989) Second Judicial District, State of Nevada. Case No. 86-5844.
WEST, G . N. (1984) The effects of auditory subliminal psychodynamic activation o n state
anxiety. Disxertation Abstracts International, 46, 319B. (University Microfilms No. DA
8504941)
WOLF,S., & PINSKY, R. (1954) Effects of placebo administration and occurrence of toxic reac-
tions. Journal of the American Medical Association, 155, 339-341.
ZAJONC,R. B. (1980) Feeling and t h k i n g : preferences need no inferences. American Psy-
chologist, 35, 151-175.
ZENHAUSERN, R., CIAIOLA,M., & POMPO, C. (1973) Subliminal and supraliminal accessory
stirnulacion and two trapezoid illusions. Perceptual and Motor Skills, 37, 251-256.
ZENHAUSERN, R., & HANSEN, K. (1974) Differential effects of subliminal and su raliminal ac-
cessory stimulation on task components in problem-solving. Perceptual a n B ~ o t o rSkills,
38, 375-378.

Accepted February 2.5, 1992