Classics in the History of Psychology

An internet resource developed by
Christopher D. Green
York University, Toronto, Ontario
ISSN 1492-3713

(Return to index)


      By Ivan P. Pavlov (1927)
      Translated by G. V. Anrep (1927)

Posted February 2001

[Classics Editor's note: Pavlov used both square and round brackets in his texts. These have been preserved but can lead to confusions as to which insertions are the author's and which are the editor's. Page numbers, reference numbers, and the occasional "sic" have been inserted in square brackets by the Classics editor. All other insertions (e.g., on p. 31) are by Pavlov. -cdg-]


Transition stages between the alert state and complete sleep: hypnotic stages.

In the last lecture an abundant array of facts was brought forward showing that sleep is nothing but internal inhibition which has become diffused continuously (i.e. without intervening fields of excitation) over the entire cortex and has descended also to some of the lower parts of the brain.

Since, as we know, the spread of inhibition is a gradual process involving first a smaller and then a greater area we should expect to find different extensities as well as different intensities of sleep or, in other words, transition stages between the fully alert state and complete sleep. Such transition stages actually exist; we had many opportunities to observe them and to produce them experimentally.

In our experiments we came across not only the usual form of sleep, which is evidenced by an absence of the normal function of the cortex and a relaxation of the skeletal muscles (closure of the eyes, drooping of the head, sagging limbs, and body limply hanging in the loops of the stand), but also a quite different form, so far, at any rate, as could be judged by the condition of the skeletal muscles.  In this form the activity of the hemispheres is also absent: all  conditioned stimuli remain without effect, and different extraneous stimuli, unless exceptionally powerful, fail to evoke any reaction. Nevertheless the animal preserves an entirely alert posture; it stands with wide open immovable eyes, head up, extremities extended, not seeking support in the loops, remaining motionless sometimes for minutes and sometimes for hours.  On changing the position of an extremity such extremity retains the new position. The flexor reflex evoked by touching the planta assumes the character of a contracture.  The presentation of food brings no reaction and the animal continues to remain quite still.  This form of inhibition was noticed only in a small number of dogs, and up to the present we are not in a position to define the special conditions of experimentation or the special peculiarities of the nervous system which are necessary for its production.  My collaborator, Dr. Rojansky, [p. 266] has made most careful observations upon the transition of dogs from the alert state to sleep, and he finds that the condition just described is present in all dogs, though usually only in a fleeting form.

It seems that the physiological interpretation of this state should not present any great difficulty: we are dealing with a complete inhibition confined exclusively to the cortex, without a concurrent descent of the inhibition into the centres regulating equilibrium and maintenance of posture (centres of Magnus and de Kleijn); in other words the animal is in a state of catalepsy. Thus in this form of sleep the plane of demarcation between the inhibited regions of the brain and the regions which are free from inhibition seems to pass just beneath the cerebral cortex.  A similar demarcation of excitable areas from areas which have undergone complete inhibition may exist also between different large areas of the cortex itself, producing what may be called a localized sleep. This form of sleep was met with frequently, and we are now able to produce it experimentally as well. On the first occasion it was observed as follows [experiments of Dr. Voskressensky]: A dog in which work had hitherto proceeded without any interference by sleep began to show signs of drowsiness -- due to its being frequently left in the stand in the experimental room for hours at a stretch without any application of conditioned or any other stimuli. Obviously the monotony of the stimulation by the constant surroundings led finally to a development of an intense inhibition involving finally the whole of the brain. The inhibitory effect of the environment became so strong that the mere introduction of the dog into the experimental room had an immediate and obvious inhibitory effect which became still more pronounced after the animal was placed in the stand.  It had to be roused up in all manner of ways to keep it from falling fast asleep before the preparations for the experiment had been completed (a matter of only a few minutes). When the experimenter left the room and closed the door in order to start the experiment from outside, and then without losing a minute began to apply one or another conditioned stimulus, the normal conditioned reflex was fully present; a normal secretion of saliva was obtained and the animal immediately took the food. When, however, after leaving the room an interval was made of 4-5 minutes before the application of the first stimulus, this stimulus now produced the following remarkable result. The conditioned secretory effect was present and [p. 267]  the salivary secretion was sharply augmented on food being presented; the animal however did not take the food, which in order to effect adequate reinforcement had to be placed in its mouth. During this time no relaxation of the skeletal muscles could be observed. When an interval of ten minutes was made after leaving the room no conditioned alimentary reflex could be obtained, and the animal was found fast asleep with a relaxed musculature and occasional snoring. Only one possible explanation of these observations suggests itself.  The inhibition must have spread in the first place only over the motor area of the cortex, so that excitation could be initiated by a conditioned stimulus belonging to any other analyser and could spread to the salivary gland but not to the muscles -- with a resulting one-sided alimentary reflex lacking its motor component.  Later the inhibition spread over the whole mass of the cortex and over the lower parts of the brain, bringing about complete sleep with a relaxation of the skeletal muscles.  In this experiment the stages of a gradually developing sleep were brought about under the influence of the protracted action of neutral stimuli upon the hemispheres, but more usually this effect appeared as the result of numerous applications within a single experiment of negative or positive conditioned stimuli, especially if, in the latter case, either the intensity of the stimulus was weak or the period of application prolonged.  The following two examples may serve in illustration:

In the first example the subject of the experiment was a dog which has already been mentioned in previous lectures (p. 231).  A tone of 256 d.v. was used as a positive conditioned alimentary stimulus, while ten neighbouring tones up and ten down the scale were differentiated [experiments of Dr. Bierman].

The second example is taken from experiments upon a dog which many exceptionally constant alimentary conditioned reflexes. [p. 268]

In this case the dissociation between the motor and the secretory response was more permanent and occurred even in experiments in which the causative inhibitory stimulus was not used.

A fresh reflex, to the appearance of a grey screen, was established and a series of experiments were performed, in each of which the new stimulus was repeated many times in succession at short intervals. It was now observed that on application of any of the old stimuli, though the conditioned secretion was often still considerable, the animal did not touch the food presented to it in reinforcement [experiments by Dr. Rosenthal].

This condition was independent of any application of the grey screen in the particular experiment and lasted for a considerable time. During the experiment the dog remained almost motionless, but there were no obvious signs of sleep. Food presented to the animal in the same stand and under precisely the same environmental conditions, but without a previous application of a conditioned stimulus, was taken with avidity.

The following chance observation belongs to the same group of phenomena. A dog which served for experiments with conditioned alimentary reflexes, and which never showed any dissociation of the secretory and motor components of the reflex, nor any signs of sleep while in the stand, was placed for the very first time in front of a large audience for the purpose of a demonstration. The unfamiliarity of the surroundings had a big effect upon the animal; it shivered slightly, and became as though spellbound. On administration of the conditioned stimulus the normal secretory effect was obtained, but the dog did not take the food, and in a relatively short time fell into profound sleep in its stand, right in front of the audience, with complete relaxation of the skeletal muscles. Evidently in this case the powerful, unusual and protracted extraneous stimulus produced at first a partial inhibition affecting only the motor area [p. 269] of the cortex; then the inhibition spread over the whole cortex and descended also to the lower parts of the brain. The experiment on the whole is similar to those by which so-called animal hypnotism is usually demonstrated.   For example, a rapid immobilization of an animal held on its back also leads to an inhibition which spreads to a varying degree in different animals.  In some cases a complete or partial catalepsy is produced (immobility of the body, but with movements of the eyes, head and neck); in others it leads to the development of profound sleep. In our laboratory this was observed on several occasions.  An extremely unruly animal, for example, vigorously resisting the preparations for the experiment, would be rapidly immobilized by a powerful grasp, associated of course with considerable mechanical stimulation, and would fall asleep in its stand almost immediately.

We see in this manner that a partial as well as a complete sleep can be produced by weak and protracted neutral stimuli, by short but vigorous stimuli, and by negative as well as positive conditioned stimuli.  I shall have an opportunity of discussing certain further details in the next lecture.

The above experiments demonstrate that the extent of the spread in the brain of the diffused inhibition can be small or great, and that there may exist different transition stages in the depth of the inhibition, or, in other words, different intensities of the diffuse inhibition (sleep).

In the eighth lecture I discussed the mechanism by which, in the case of simultaneous conditioned stimuli, a stimulus from one analyser is overshadowed by a stimulus from another, and the suggestion was put forward that this overshadowing might be dependent upon the different strengths of the stimuli belonging to the different analysers (p. 141).  Experiments which have since been performed entirely uphold this suggestion.  When we intentionally produced a considerable change in the strength of our usual conditioned stimuli, the auditory being made weaker and the others either remaining unchanged or being made stronger, there was a definite reversal of the relations previously obtained, the auditory stimuli now participating in the stimulatory compound to a smaller extent than the other stimuli, i.e. on isolated application of an auditory stimulus a much smaller effect was obtained than on application of any other stimulus belonging to the compound.  The following are some of the experiments: [p. 270]

In one dog a compound simultaneous conditioned stimulus consisted of a tactile and an auditory component, the auditory being considerably weakened. The compound stimulus, when well established, gave 4-4 1/2 drops of saliva during 20 seconds' isolated action. When used separately the auditory component gave a secretion of 1-1 1/2 drops and the tactile 2 1/2 -5 drops [experiments of Dr. Rickman].

In another dog the compound simultaneous conditioned stimulus was made up of a 100 candle-power lamp together with the sound of a musical tone which was considerably damped.  The compound stimulus when fully established gave 7-8 drops of saliva during 30 seconds; the visual stimulus applied singly gave 5 drops, and the auditory gave 2 1/2 drops.  In an exactly similar manner a thermal cutaneous stimulus of 0? C., which was employed with a very weak one to form a compound simultaneous conditioned stimulus, gave when applied singly a much greater effect than the tone [experiments of Dr. Gantt and Dr. Koupalov].

Thus we see that the difference in the intensity of the reflexes evoked by the various conditioned stimuli belonging to the different analysers is determined by the strength of the stimulus and not by any functional difference in the nervous elements of the analysers. These experiments give us a method of comparing the intensity of stimuli which belong to different analysers.

Bearing these facts in mind we can begin to study the different stages through which the diffused inhibition passes in its development.  The starting-point for these investigations was provided by a case of a pathological state of the nervous system which had been brought about experimentally by means of a purely "functional (non-surgical)  interference."   Experimentally  produced  pathological states of the nervous system will be dealt with fully in succeeding lectures; in the present lecture I shall describe only the experiment which induced us to pursue the further investigation of normal animals.

Positive conditioned alimentary reflexes [experiments by Dr. Rosenkov] were established to the sound of a whistle, beats of a metronome, rhythmic tactile stimulation of the skin at a rate of 24 per minute, and flashes of an electric lamp; several negative reflexes were also established by differentiation, including one to tactile stimulation of the same skin area at the rate of 12 per minute. The following table gives the figures for the normal effect of the positive stimuli:   [p. 271]

On the basis of the previous discussion we may take the strength of the stimuli in order from strong to weak as whistle, metronome, tactile stimulation and lamp.

The experiment now proceeded as follows: In between the different positive stimuli the differentiated tactile stimulus of 12 per minute was introduced, being applied during 30 seconds and followed without any interval by the positive tactile stimulus of 24 per minute which was also continued for 30 seconds and then reinforced as usual.  This seemingly small factor produced an extraordinary effect.  On the day following this experiment and on the succeeding nine days all conditioned reflexes had disappeared excepting only for a very occasional small secretion.  This period was followed by a series of definite successive changes in the conditioned activity of the brain. The first of these extremely peculiar changes is illustrated by the next experiment.

The experiment shows exactly the reverse of what was observed during the normal state of the animal. The strong stimuli have either no effect or only a very small one; the weak stimuli have a greater effect than normal.  All positive stimuli were, of course, reinforced.  This state of the cortex we called the paradoxical [p. 271] phase. The paradoxical phase in this dog continued for fourteen days and was then succeeded by the following change:

This was called the phase of equalization, since all the stimuli became equal in their effect. The phase of equalization lasted for seven days and was then succeeded by still another phase during which the effect of stimuli of medium strength was greatly increased; the effect of the strong stimulus was slightly diminished, while the weak stimulus had no effect.  After seven days more, all the reflexes had returned to their normal value. In the succeeding experiments, on the same problem, in order to be quite certain, we used different intensities of one of the positive stimuli.  The results obtained were exactly comparable with the results of the previous experiments. It thus became obvious that the difference in the reaction to stimuli in all these different phases is determined by the relative strength, of the stimuli.

In the manner just described was secured the first direct evidence that the cellular structures of the cortex undergo a series of definite stages of transition between complete inhibition and normal excitability, stages which are divulged by the peculiar reactions of the cortical elements to the stimuli of different strengths.  After the study of these transition stages in this obviously pathological state, the question arose whether the same stages would be found normally during the transition from the alert state to sleep and the reverse.  It was thought probable that the pathological case just described consisted only in an exaggeration and prolongation of events which in the normal animal were transient and not so evident, just as was the case with catalepsy.  Special experiments conducted in this direction led to a definitely positive result.  The following are illustrations:

Twenty neighbouring tones had been differentiated from the tone acting as a positive stimulus.  This dog had also, among many  [p. 273] others, two positive conditioned reflexes, differing greatly in intensity, to a weak and to a loud crackling sound.  The following table gives the normal intensity of these two reflexes:

The actual experiments proceed as follows. By repeated application  of the differentiated tones the animal is rendered definitely drowsy; the weak crackling sound is now applied. The secretory is absent.   The dog awakens during the reinforcement with food, which it begins to eat. The next application of the weak crackling sound evokes a secretion which is yet small.  The reflex is again reinforced.  A third application of the weak crackling sound produces a normal, or in some cases even a supernormal, secretory effect, and the reflex is again reinforced.  The strong crackling sound is applied next;  its effect is either less than or equal to the last effect of the weak sound.  It is only somewhat later, when the alert state has been fully recovered, that the strong crackling sound evokes its full normal effect, and that the normal quantitative relations between the two reflexes become restored.  The following gives one of the actual experiments:

[p. 274]

In some experiments repetition of these stimuli instead of leading to a temporary predominance of the effect of the weaker stimulus resulted only in an equalization of the effects of the strong and weak crackling sounds.  Evidently during the gradual dispersion of sleep under the action of repeated brief feedings the cortical elements pass through the paradoxical phase and the phase of equalization.  It follows that these experiments are exactly comparable to the pathological case which was previously described, excepting that the change which took under normal conditions a few minutes required in the pathological case many days.

In another dog a slight drowsiness developed on account of too prolonged experimentation.  This was accompanied by a complete obliteration of the differences in the intensities of the reflexes to the different stimuli, so that all the positive conditioned reflexes now became equal. With the help of injections of a suitable dose of caffeine the dog was brought back to its usual condition of wakefulness, and with this all the normal relations between the intensities of the different conditioned reflexes returned.  Both experiments [by Dr. Zimkin] are given below:

On the following date the day the animal received subcutaneously 8 cc. of a 2% solution of caffeine eighteen minutes before the experiment.  At the time of the experiment the animal was fully alert.

  [p. 275]

In the animal which was previously mentioned in this lecture as showing dissociation of the secretory and motor reactions, it was often observed that during the period of this dissociation the weakest conditioned stimulus (the lamp) was the only one which evoked on some occasions a strong salivary reflex, sometimes even bringing about both reactions -- a normal secretion and motor response, and acceptance of the food on reinforcement.  It is thus seen that a paradoxical phase could be observed also in these cases of a limited extent of diffusion of the inhibition [experiments by Dr. Rosenthal] .

A further and quite peculiar condition was observed in some cases of intense drowsiness which fell just short of changing into complete sleep.   When positive conditioned stimuli had nearly lost their effect, well-developed negative stimuli, on the other hand, acquired definite excitatory properties.  The following is an example of such an experiment by Dr. Shishlo:

Positive conditioned alimentary reflexes were established to tactile stimulation of the shoulder and of the thigh, and to a thermal cutaneous stimulus of 45? C.; a very constant negative conditioned stimulus was also established to a tactile stimulation of a definite skin area on the back.  The effect of the positive tactile stimuli ranged normally from 15-18 drops during one minute.  The thermal conditioned stimulus began relatively soon to induce drowsiness and sleep.  The experiment to be described commenced with an application of the thermal cutaneous stimulus which led to drowsiness. experiment then proceeded as follows:

A similar conversion of negative stimuli into positive ones was also on several occasions observed in pathological conditions. This effect is given the name of the ultra-paradoxical phase.

It thus becomes evident that during the transition from the alert  [p. 276] state to complete sleep the hemispheres pass through several different stages. Since sleep is nothing but a widely distributed interned inhibition, we should expect at least some of these stages to appear during the ordinary inhibitory after-effect, which was discussed at length in the earlier lectures upon internal inhibition.  So far only one case has been investigated, namely, conditioned inhibition, and this would appear to realize our expectation [experiments by Dr. Bikov]:

Five positive conditioned reflexes were established -- to a metronome, a loud tone, the same tone damped down, the appearance of a disc of cardboard, and tactile stimulation of the skin. A conditioned inhibition was developed to a combination of a sound of bubbling with the action of the tactile stimulus. The mean figures of the secretory effect of the five positive stimuli, averaged from a great number of experiments, were in the above-mentioned order of the stimuli -- metronome 22, loud tone 18 1/2, soft tone 16 1/2, disc 13 1/2, and tactile stimulus 10 drops during 30 seconds.  The conditioned inhibition having been firmly established, all the conditioned stimuli in turn were tested 10 minutes after the application of the inhibitory combination.  The metronome gave 16 1/2, the loud tone 16, the damped tone 20, and the disc 18 drops. Taking into consideration the possible interferences of irradiation of inhibition and of induction, the only point of importance in the present connection is that the effect of the weaker tone was considerably above normal, while that of the stronger tone was below normal.  This reversal in the effect of the strong and weak tones can be regarded as evidence of a paradoxical phase, since the tone was the same, differing only in strength, and both stimuli therefore were obviously related to the same point of the cortex.  This investigation is at present being continued with other types of internal inhibition.

In the lecture upon mutual induction a suggestion was made that external inhibition might be due to negative induction, i.e. to an inhibition which is induced in cortical areas neighbouring on the area of excitation.  Expressed in another fashion, it was suggested that the intimate mechanism underlying external inhibition is identical with that underlying internal inhibition.  It was hoped to test this theory by determining whether external inhibition causes similar changes in the reactions of the cortex to those which have just been described in the case of internal inhibition.  For the purpose of this investigation a stimulus was needed which would produce a [p. 277] protracted effect of external inhibition, and use was made of the introduction into the animal's mouth of rejectable substances which, as was mentioned previously, produces a prolonged after-effect.  The experiments were performed upon two dogs, both of which had well-established alimentary conditioned reflexes.

In the first dog [experiments by Dr. Prorokov], after introduction of a solution of sodium carbonate, strong and weak conditioned stimuli were tested immediately on termination of the secretion due to the alkali itself.  It was found that at first all the conditioned reflexes were inhibited to the same extent, but that within the next 15-20 minutes the reflexes to the weak stimuli returned to normal or even exceeded the normal value, while the strong stimuli were either equal in effect to the weaker ones or even gave a considerably smaller effect. In the experiment given below a solution of sodium carbonate was introduced into the dog's mouth at 9.41 a.m.

Under the usual conditions without administration of the alkali the effect of the buzzer was about 8 drops during 30 seconds, while the effect of the tactile stimulus was 4 drops during 30 seconds.

In a second dog [experiments by Dr. Anokhin] there were, however, somewhat different results.  After the introduction of the rejectable substance into the dog's mouth and the termination of the resulting secretion, all the conditioned stimuli, when tested at frequent intervals up to the end of the experiment, showed an equalization in their effect.  Concurrently with this there was observed as the experiment continued a step-like diminution in the strength of the reflexes.  In control experiments performed previously the reflexes were proportional to the strength of the stimuli, the stronger buzzer giving the largest effect and the lamp the smallest.  In the following example a solution of sodium carbonate was introduced into the dog's mouth at 11 a.m., and the resulting secretion of saliva continued for ten minutes.  [p. 278]

Although the results obtained in the two dogs would appear to corroborate the suggestion that internal and external inhibition are fundamentally one and the same process, yet the complexity of the problem necessitates a repetition and greater variation of the experiments with more critical attention to other possible interpretations of the results.

In the course of our investigation we became greatly interested in the effect upon conditioned reflexes of different narcotics in the first stages of their action, in complete narcosis, and again during the period of recovery.  Urethane and chloral hydrate were used for this purpose.  In the case of the action of narcotics as compared with the effect of inhibition a different sequence of events was observed: there was a gradual weakening of all conditioned reflexes, the weak conditioned stimuli naturally becoming ineffective before the strong ones.  This state was given the name of the narcotic phase.   The following experiment is taken from a research by Dr. Lebedinsky:

Positive conditioned reflexes were established to loud buzzing, metronome, weak buzzing, tactile stimulus, and intermittent flashes of an electric lamp. With regard to the intensities of their effect the stimuli followed in the order given. The animal, after being placed in the stand, received at 10.9 a.m. two grammes of chloral hydrate dissolved in 150 cc. of water in the form of an enema. The experiment proceeds as shown in the table on the opposite page.

We thus see that with the development of narcosis the effect of all the stimuli progressively diminished, and on return to the alert state the stimuli progressively recovered their normal conditioned effect.  The only exception, out of the twenty stimuli, was presented by the weak buzzing sound which at 11.53 a.m. produced an abnormally large effect.  [p. 279]

Thus in different healthy animals under different conditions there were found many different phases of transition in the reactions of the cortex to conditioned stimuli.  An obvious question arises as to how far all these different phases, including also the narcotic phase,

appertain to every single animal under the usual conditions of life. In investigating this problem we were fortunate to have at our disposal an unusually reactive type of animal, the special features of which will be discussed in the succeeding lectures.  (This dog was used for the experiments described towards the end of the fourteenth lecture, p. 247).  Given constant conditions, the dog was remarkable for the constancy of its highest nervous activity as expressed in the  [p. 280] form of conditioned reflexes, and fully merited its nickname of " an animated instrument."  The dog had ten different conditioned reflexes.  There were six positive ones -- to a buzzer, metronome, whistle, increase in illumination of the room, appearance of a small toy horse ; and four negative ones -- to a different rate of metronome, a diminution in the illumination of the room, the appearance of a square, and the appearance of a toy rabbit approximately of the same size and colour as the toy horse. For some time prior to the following experiments the buzzer had not been used, while out of the differentiated stimuli the metronome had been employed practically exclusively.  The auditory stimuli usually, and in the early experiments always, gave a secretion 30 to 50% greater than that given by the visual stimuli.  After two years' work in the laboratory with this dog the positive conditioned reflexes showed some tendency to diminish and began to vary in their relative intensity, as frequently happens in the case of continuous and prolonged use of the same conditioned stimuli. During this state of the animal we could find all the definite transition stages of the cortical activity which were described earlier in this lecture as different phases of the progressive diffusion of the inhibitory process over the hemispheres. Each of these phases either lasted during the whole course of a single experiment or else, under the action of different influences produced experimentally, changed into some other phase [experiments by Dr. Speransky].  The only phase which could not be observed in this dog was the ultra-paradoxical, but the conditions were not such as to favour the appearance of this phase as the animal never became very drowsy. The experiments taken from different periods of the work are shown on the opposite page.

When the reflexes deviated very much from normal they were strengthened and corrected by the method discussed in the fourteenth lecture, namely, by abbreviating the interval between the beginning of the conditioned stimulus and its reinforcement. The spontaneous transition from one phase to another in the two last experiments was most probably due to repeated reinforcement. We had, however, two special methods at our disposal by the use of which we could produce an immediate interchange of phases. One of these methods consisted in the application of the completely differentiated inhibitory rate of the metronome. Most probably the effect of this inhibitory stimulus was due to concentration of the diffuse inhibition, or to an induction of the antagonistic process of excitation.  [p. 281]

[p. 282]

The second method consisted in the application of an extra stimulus, viz. the entry of the experimenter into the animal's room.  The following experiments serve as examples of the effect of each method :

The phase of complete inhibition with absence of the secretory and motor reactions is transformed in the above experiment first into the paradoxical, and then into the normal, phase by the application of the inhibitory rate of the metronome.

The presence of the experimenter in the room with the dog immediately changed the phase of complete inhibition to normal.

The question whether all these different stages in the activity of the hemispheres can be arranged in a definite order, and if so in what kind of order, must remain for the present entirely open. A consideration of all the experiments at our disposal shows that the [p. 283] sequence of the different phases was fairly variable, and it is not clear, therefore, whether the different states of the hemispheres in different animals are strictly successive or whether they may occur as parallel events.  Only further experimentation can explain why a given phase undergoes transition directly into one or another of the remaining phases.

It is hardly possible to doubt that all these different states of the hemispheres bear a strong resemblance to the different stages of what is generally known as hypnotism. The relation between the experimental results described here and hypnotism as observed in man will form the subject of the last lecture.