RU2293518C1 - Method for estimating sincerity-insincerity of speaking person - Google Patents

Abstract

FIELD: medicine.

SUBSTANCE: method involves using triple estimation of speaker emotional and psychological properties and states from single video recording fragment by applying psychological scaling approach. The first stage involves giving estimation based only on listening to speaker voice with video image being switched-off. The second stage involves giving estimation based only on video image with the voice being switched-off. The third stage involves giving estimation based on speaker image and soundtrack. Correspondence coefficient of mean values audio and video estimates in scores given by experts at the first and the second stage of speaker study based on above emotional and psychological properties and states of the speaker. Sincerity-insincerity is estimated by calculating average estimates of all emotional and psychological properties and states of the speaker equivalent to notions of sincerity and insincerity.

EFFECT: high reliability of estimation method.

3 cl, 7 dwg

Description

The proposed method relates to the field of psychology, more precisely, social psychology (A.L. Zhuravlev), psychophysiology of speech and non-verbal communication.

The sincerity of the speaker is widely known in the everyday life of interpersonal relations and, at the same time, is a concept far from studied by science. As a rule, sincerity is interpreted as a positive property of the speaker, as a kind of ethical norm of speech behavior. In the dictionaries of V. Dahl and S.I. Ozhegova: “Sincere - it means truthful, frank, honest, conscientious, expressing genuine feelings, sincere, unprincipled, etc.".

The connection of sincerity with truthfulness does not mean, however, the identity of these concepts in the formal logical sense [V.V. Signs. Psychology of understanding the truth. St. Petersburg, 1999, pp. 103-104]. The speaker may be inaccurate in his judgments, but at the same time, if he is sincere, then he says what he thinks and feels, i.e. his statement is fully consistent with his thoughts and feelings. Thus, the truthfulness of a sincere person does not consist in the fact that he is always objectively right (i.e., expresses the objective truth), but that he is subjectively right, i.e. not contrary to their thoughts and feelings about the subject of conversation. Sincerity is the conviction of one’s truthfulness. Sincerity gives strength and persuasiveness to speech, inspires confidence in the speaker and his words. Insincerity reduces the persuasiveness of speech, even if it is logically flawless. [A.F. Horses. Judicial and political eloquence // About oratory. M., 1963, pp. 155-191].

As for insincerity, this concept is almost equivalent to the concept of "untruth" or "false", because in the case of insincerity, a person does not say what he thinks and feels.

Along with the recognition of sincerity as the undoubted dignity of a person, the speaker’s insincerity, that is, false [see Signs, cit. ed. 1999] and, not only in Russia, but also in other countries, as M. Montaigne wrote about: "... To lie and constantly break the word of the French is by no means a vice; for them it’s the same as the way of talking" [M. Montaigne. About the exposure of lies // Experiments. Ed. 1998, book 2, pp. 801-806].

The paradox of the situation is that there has been a “peaceful coexistence” of sincerity and insincerity in society in all historical times to this day, with a general condemnation of the latter, honesty and deception, truth and falsehood. This is manifested not only in the fact that certain restrictions are imposed on the sincerity of the speaker (“Truth to Speak - Annoy Yourself”, V.I. Dahl), but also in the existence of numerous varieties of legalized forms of lies. These include: a lie in the name of salvation (for example, a terminally ill, or the use of a placebo), a lie to the benefit of the state and society (which Plato also allowed), a lie in diplomacy ("words were given to us to hide our thoughts," said Talleyrand) , a lie in military affairs (disinformation of the enemy in the interests of victory - "military trick"), in operations of special services ("operational combination"), in business (corporate lie in the interests of the company, AN Tarasov), finally, a lie dear to many hearts in the form of flattery, friendly practical jokes, harmless lies, far from harmless slander, etc. etc.

In connection with all of the above, the problem of recognizing the sincerity-insincerity of the speaker arises.

The phenomenon of sincerity-insincerity is detected in the process of communication, expressed as verbal verbal (word), and non-verbal means (intonation of voice, facial expression, posture). In this regard, research on the issue of assessing the sincerity of a speaker is divided into two categories, respectively.

A number of researchers follow the path of analysis of verbal, semantic, logical analysis of the actual textual part of the speaker’s statement [A. Myagkov Experimental strategies for diagnosing and measuring the sincerity of respondents // Sociological studies. M., 2003, No. 2, pp. 115-125]. The unreliability of the verbal methods commonly used in social research (questionnaire survey of the population) is determined by the strong dependence of the responses of the subjects on the degree of social desirability. Based on personal data, you can never be sure - sincerely, i.e. truthfully, the person answered or evasively, or wrote an absolute lie, especially in sensitive issues related to assessing the political situation in the country, intimate issues of a person’s life, and even more so - the respondent’s involvement in illegal acts.

Other researchers focus on non-verbal signs of lies, which for several reasons is considered more reliable due to the involuntary nature of non-verbal signs of lies. The famous American specialist in lie detection, P. Ekman, for example, has up to 114 features of non-verbal behavior in insincere statements (P. Ekman. Psychology of lies. St. Petersburg, 2000, pp. 235-242). However, the use of this "catalog of lies" in practice encounters great difficulties for the following reasons.

1. Despite its considerable volume, it does not cover the whole variety of signs of lies in view of the significant individual differences in the ways of expressing lies.

2. He does not evaluate the severity of the signs of lies, although the severity of these signs even in the same speaker, and even more so in different people, can fluctuate within very significant limits.

3. Ambiguities on what to pay attention to the observer (voice, face, hands of the speaker).

4. Significant differences among observers in the ability to recognize lies according to the features proposed by Ekman.

5. Ekman's method, aimed at identifying lies, practically leaves aside the assessment of the sincerity of the speaker.

Attempts to determine the speaker’s insincerity (lies) by the characteristics of his speech or non-verbal behavior have been undertaken by other researchers, however, the problem as a whole remains unresolved. “Attempts to create effective technologies for teaching lie recognition,” write Yu.M. Zhukov and D.V. Khrenov, “immediately make obvious the insufficiency of our knowledge about the laws of manifestation of insincerity. A simple question:“ But what should one pay attention to identify false messages? "there is essentially no clear answer. Information on differences in the distribution frequencies of various verbal or nevrebale behavioral reactions does not give much for constructing clear formulations" (Yu.M. Zhukov, DV Khrenov. Methodical analysis of the study IAOD insincerity // World of Psychology, №3, 1999 str.219-230).

The most famous and trustworthy around the world, according to some experts, is a method for determining the insincerity (lies) of a speaker using a polygraph or a lie detector (V. A. Varlamov. Lie detector. M., 2004). The specified method is based on the occurrence of involuntary emotional reactions in the offender upon presentation of questions related to the offense and registration on the polygraph of involuntary autonomic reactions caused by emotion: pulse change, breathing, sweating, tremor of the hands, etc. It is assumed that persons not involved in the offense under investigation, such emotional-vegetative reactions do not occur.

1. The main disadvantage of a lie detector, which causes a significant number of errors in suspects, is that the above emotional-vegetative reactions can occur not only in those responsible for offenses, but also in innocent people. In the latter case, they arise due to a person experiencing the danger of undeserved punishment (the so-called "Desdemonna effect", the fear of which, caused by Othello's suspicion, he took as proof of her guilt and killed his innocent wife).

2. On the other hand, people who are really guilty of an investigated offense and are familiar with the method of detecting lies on a polygraph often learn to suppress the manifestations of emotional-vegetative reactions through special training or taking pills that reduce neuropsychic reactivity (V. A. Varlamov , cit. ed., pp. 308-318). These serious shortcomings of the polygraph caused the lie detector to be convicted by many undeservedly convicted people, of which only a small part were acquitted in connection with the discovery of the true culprits of the offense (P. Ekman, op. Cit., Pp. 134-135) .

3. A significant drawback of the polygraph is its limited use, since there are a number of categories of people who cannot be examined by a lie detector, the need for the written consent of a suspect to be examined by a lie detector, the complexity of the procedure for interpreting the data that require a qualified operator (verifier), and the high cost of the hardware complex of the polygraph the inability to conduct a polygraph examination of persons speaking on television, at various meetings, by video recording, etc.

4. The disadvantage of a lie detector is the inability to assess the degree of sincerity of the speaker, because this apparatus is narrowly specialized for detecting lies.

The purpose of the proposed method is to increase the accuracy and reliability of the assessment of sincerity-insincerity, the possibility of using it as a supplement to the examination on a polygraph and as an independent method.

The proposed method is based on the idea, confirmed by preliminary experiments of the authors that with a sincere statement, the speaker involuntarily forms a coordinated complex of all forms and elements of non-verbal behavior. For example, with sincere joy, not only the intonation of the voice speaks about this emotional state, but also the corresponding facial expressions, and posture, and gestures. With sincere joy, even elements of facial expressions turned out to be consistent; joy is evidenced not only by a smile, but also by the corresponding expression of the eyes, they also "laugh."

With insincere, feigned joy, when a person is not internally happy, but tries to imitate joy, such a harmonious, coordinated complex of all elements of non-verbal behavior is not formed. For example, an insincere smile is expressed in an unnatural stretching of the corners of the mouth, but the eyes remain cold or, moreover, can even express hostility (a “prickly look”). The false “joyful” intonation of the voice may not correspond to a not joyful expression on the face or contradict a far from joyful pose, gestures, etc.

The voice itself, like the appearance of the speaker, contains a number of emotionally expressive means. This is the melody of speech (changing the pitch of the voice over time), the dynamics of changing the strength of the voice over time, the temporal characteristics of speech and the timbre of the voice, i.e. the color of the sound, depending on its overtone composition. These elements of intonation can be in agreement with each other (with a sincere statement) or contradict each other (with insincerity).

Observations and experience show that almost every normal person in the process of constant communication with people develops the ability to determine a person speaks sincerely or insincerely, precisely by the degree of correspondence of the indicated elements of his non-verbal behavior, i.e. a kind of "psychological lie detector" is being formed. Naturally, the degree of development of this ability in different people is different, because depends on social experience, age and observation of a person.

But even with sufficient experience, some details may slip away from the observer from the speaker’s complex of non-verbal behavior, indicating that he is speaking insincerely, i.e. lying. This is because the observer, watching the speaker’s facial expressions (and besides the speech content!), May not pay attention to the intonation of his voice, which contradicts facial expressions, or gestures that do not correspond to facial expressions, etc.

Errors in the assessment of sincerity are also explained by the fact that during speech, elements of non-verbal behavior - intonation, facial expressions, posture, gestures - are dynamic, quickly replace each other, and therefore some fleeting details indicating the speaker’s insincerity can slip away from the observer’s attention, and all the more so that in the process of the same social experience, a considerable number of speakers develop the ability to hide their insincerity by a skillful imitation of sincerity.

In this connection, the main idea of the proposed method is to increase the reliability of assessing sincerity and insincerity by increasing the accuracy of the observer’s assessment of the degree to which all elements of the speaker’s non-verbal behavior are consistent: voice intonation and expressive movements (gestures, posture, facial expressions).

This goal is achieved in that the determination of the sincerity, as well as insincerity of the speaker is carried out according to the results of a three-time evaluation of a number of his emotional and psychological properties and conditions by expert observers on the same fragment of the speaker’s video recording with a duration of at least 40-60 s, taking into account the correspondence coefficient (To _acc. ) Auditory and visual assessments of sincerity and insincerity, based on the degree of correspondence of the intonation of the voice with the whole complex of expressive movements (facial expressions, postures, gestures).

At the same time, at the first stage, evaluations are made only on the basis of listening to the speaker’s voice when his video signal is off, i.e. only on the basis of intonation-timbre, tempo-rhythmic, loud features of the voice and their changes in time.

At the second stage, assessments are made only from the video image of the speaker with the sound of his voice turned off, i.e. based on observation of the speaker’s facial expressions, posture and gestures and their changes in time.

And at the third stage, the assessments are already given according to the normal voiced video image of the speaker.

At all stages of the speaker’s examination, assessments of sincerity and insincerity are given by the degree of their severity according to the method of psychological scaling (see Psychology. Dictionary. M., 1990, pp. 48-449) on a five-point scale or other scales, and in combination with others, semantically ( i.e., in meaning) the opposite psychological properties and states of the speaker, including properties that are equivalent or similar in meaning to sincerity and insincerity, and the speaker’s possible states, namely: sincerity-insincerity, frankness-secrecy, clusive-weakness, candor-deceit is generally not less than 3, 4 pairs of similar candor-insincerity properties and states of the speaker (see. synonyms candor in the above-cited dictionary by Dahl and Ozhegova SI).

The speaker’s properties and possible states of the speaker also include a number of his possible emotional states, both general and specific, also in the form of semantically opposite pairs: emotionality-indifference, optimism-pessimism, peacefulness-aggressiveness, peace, anxiety, courage-fear, sensitivity-rigidity, also generally not less than four to five pairs.

Thus, the list of subject-matter emotional-psychological properties and possible states of the speaker as a whole includes at least 16-20 of his properties and conditions.

The correspondence coefficient of the speaker’s auditory and visual assessments (K _acc. ) _Is determined by the results of average assessments for a group of expert observers by calculating the Spearman rank correlation coefficient (see E. Sidorenko. Methods of mathematical processing in psychology. St. Petersburg, 2001, p. 207- 223) by comparing two series of average ratings of all the emotional and psychological properties and states of the speaker obtained in the first and second stages of the examination.

The Spearman correlation coefficient calculated in this way is taken as the value of the correspondence coefficient (K _resp ) of the auditory and visual assessments of the speaker’s non-verbal behavior, and according to the main idea of the proposed method, the psychophysiological measure of the speaker’s sincerity.

In the future, the average values of the sincerity and insincerity of the speaker are calculated according to the estimates of all expert observers given by them at the third stage of his examination, and the sincerity is determined by calculating its average total score based on the assessment of all those psychological properties and conditions of the speaker that are close to the concept of "sincerity ", including: sincerity, frankness, persuasiveness, honesty and other, in general, at least three to four characteristics. The result obtained is multiplied by the previously calculated correspondence coefficient (K _acc. ) And the final sincerity value is expressed as a percentage in relation to the maximum score of its assessment.

The insincerity is determined in a similar way, i.e. according to the results of averaging the estimates of all expert observers given by them at the third stage, by calculating the average score of the insincerity and related psychological properties and conditions of the speaker, including: insincerity, secrecy, unconvincing, deceit, and others, not less than three or four characteristics, and thus obtained the total average value of insincerity is multiplied by the value: (1-K _resp. ); where K _acc - the coefficient of conformity, and express the final result as a percentage in relation to the maximum point of insincerity.

In order to increase the accuracy and reliability of the sincerity assessment, a group of expert observers is formed from among persons (at least 10 people) with an emotional hearing of at least 80% according to the V.P. test. Morozov's emotional hearing (RF patent No. 2221487). In addition, experts should know the basics of expressive human movements in the volume of at least popular publications: J. Nirenberg and G. Calero. Reading a person is like a book. M., 1990; A. Pease. Body language. N. Novgorod, 1992 and others.

Assessment of sincerity-insincerity and other emotional and psychological properties and possible states of the speaker, expert observers register in special forms presented in figures 1, 2, 3, the results of which are to be processed according to the above method.

Along with the advantage of the blank method (ease of use), it also has a significant drawback, which is associated with the complexity of computing the average values of many indicators.

In order to simplify the procedure for processing the results, as well as improve accuracy (since the blank methods conceal the probability of inevitable errors in the preparation of matrices, mathematical operations, etc.), it is advisable to use a computer method. Its essence lies in the fact that each of the observer experts is provided with a computer with DVD software that allows the expert in his individual mode to assess the sincerity-insincerity of the speaker, i.e. stop the video frame if necessary, return for repeated playback, partially slow down the playback, etc. They unite the personal computers of experts in a single local network, reduce all information to the main computer, which, using specialized software, processes and issues the final results of the speaker’s sincerity-insincerity assessment in real time.

Thus, increasing the accuracy and reliability of determining the sincerity-insincerity of the speaker by the proposed method is carried out by the following means.

Firstly, by three-stage perception by expert observers of the same video clip of the speaker;

secondly, by introducing a correction factor (K _acc. ) to the assessment of sincerity-insincerity, reflecting the degree of conformity of auditory and visual evaluations of expert observers;

thirdly, an increase in the number of properties and conditions of the speaker to be evaluated due to characteristics close to sincerity and insincerity;

fourthly, by applying quantitative methods of psychological scaling and semantic differential (semantically opposite pairs);

fifthly, by increasing the number of expert observers to at least 10 people with emotional hearing according to the V.P. test Morozov (patent No. 2221487) at least 80% and familiar with the basics of expressive movements of the speaker;

sixthly, using computer methods for obtaining and processing the results of real-time speaker sincerity-insincerity assessment.

In the future, the proposed method is illustrated by specific examples of its application, the attached form forms for expert observers and graphs explaining the results of calculating the sincerity-insincerity of the speakers.

Figure 1, Figure 2, Figure 3 - forms of forms of expert observers for use in the first, second and third stages of the examination of the speaker (s), respectively.

Figure 4 - comparative results of average auditory and visual assessments of emotional and psychological properties and conditions of the speaker No. 9.

Legend.

Vertical - average assessments of expert observers in%.

Horizontal - emotional and psychological properties and conditions of the speaker No. 9, ranked for auditory assessments according to their severity from minimum (left) to maximum (right).

Line 3 - auditory scores.

Line 4 - visual assessments.

Figure 5 - comparative results of average auditory and visual assessments of emotional and psychological properties and conditions of the speaker No. 4.

Legend.

Vertical: the value of the estimate in%.

Horizontal: emotional and psychological properties and conditions of the speaker No. 4, ranked for auditory assessments according to their severity from minimum (left) to maximum (right).

Line 1 - auditory scores.

Line 2 - visual assessments.

Figure 6 shows the relationship between the correspondence coefficient (K _acc. , Line 5), average assessments of sincerity excluding K _acc. (Line 6) and taking into account K _acc. (Line 7) for a group of 11 speakers.

Legend.

Vertical: on the left - the average assessment of sincerity (%), on the right - the value of the coefficient of compliance (K _mat .).

Horizontal: speaker numbers. The sequence of arrangement of numbers of speakers is ranked by the value of the final assessments of sincerity (i.e., taking into account K _respectively ) from maximum (left) to minimum (right).

FIG. 7 - the relationship between K _resp (line 5), average insincerity estimates excluding K _resp (line 8) and taking K _resp (line 9) into account for a group of 11 people.

Legend.

Vertical: on the left - the average assessment of sincerity (%), on the right - the value of the coefficient of compliance (K _mat .).

Horizontal: speaker numbers. Sequence numbers speaking location ranked highest final insincerity changes (i.e., with the K _{respectively)} of the minimum (left) to the maximum (to the right).

The proposed method for assessing the sincerity-insincerity of the speaker is as follows.

At least 10 people are formed as a preparatory stage for a group of expert observers. The main criterion for the selection of expert observers is that each of them has an excellent emotional hearing, i.e. not less than 80% by test V.P. Morozov's emotional hearing (VP Morozov. Method for assessing a person’s emotional hearing. Patent No. 2221487). Studies have shown that the presence of good emotional hearing testifies to a person’s belonging to an artistic type of person, his ability to recognize emotions and other psychological properties and conditions both in terms of his voice and appearance of the speaker (Artistic type of person. Complex studies. M., 1994; D.V. Lyusin, Modern Views on Emotional Intelligence // Social Intelligence, Theory, Measurement, Research / Edited by D.V. Lyusin and D.V. Ushakova, M., 2004, pp. 29-36 ) Experts should also know the basics of human expressive movements in the volume of at least popular publications: J. Nirenberg and G. Calero. Reading a person is like a book. M., 1990; A. Pease. Body language. N. Novgorod, 1992 and others.

The preparatory phase also includes the preparation of the speaker’s video fragment of interest to the researcher, which should be at least 40-60 s from the prerecorded video tape of the speaker from the TV screen or in any other conditions. In the case of an examination of several speakers, the video fragments are copied to a separate video tape or DVD-ROM with the numbers of the speakers indicated by the voice of the speaker preceding the next video fragment.

Before the start of the survey, each expert observer is given three forms for entering assessments of sincerity-insincerity and other emotional and psychological properties and states of the speaker. Forms of forms are presented in the appendix (Fig. 1, Fig. 2, Fig. 3, respectively, for the first, second and third stages of the examination of the speaker).

Each of the forms contains a list of the emotional and psychological properties of the speaker to be evaluated, presented in pairs in the form of semantically (i.e. in meaning) opposite pairs, including emotionality-indifference, optimism-pessimism, sincerity-insincerity, peacefulness-aggressiveness, persuasiveness- unconvincing, calm-anxiety, frankness-secrecy, courage-fear, frankness-craftiness, sensitivity-rigidity.

According to a brief instruction that is given on each of the forms, the observer expert should evaluate each of the speaker’s properties and states on the form in points, marking the corresponding number (cross out or circle) in the series of numbers 4 3 2 1 0 1 2 3 4, located between the names of semantically opposite properties and psychological states of the speaker.

Since the assessed qualities are opposite, then, of course, only one figure is noted, corresponding to the degree of expression of that quality, which, according to the expert, is expressed in one way or another in the speaker, and the opposite quality is thereby automatically evaluated as a zero point. In case of non-expression of either quality, the pairs mark the number "0".

The advantage of this kind of alternative assessments is that the expert, putting down 10 assessments, thereby evaluates the 20 emotional and psychological properties and conditions of the speaker, presented in the expert’s form. This method of "semantic differential", as it is often called, is considered one of the preferred and accurate methods for quantifying the psychological properties and conditions of a person.

Assessment of sincerity-insincerity is carried out in three stages.

At the same time, at the first stage, evaluations are made only on the basis of listening to the speaker’s voice when his video signal is off, i.e. only on the basis of intonation-timbre, tempo-rhythmic, loud features of the voice and their changes in time.

At the second stage, estimates are given only by the video image of the speaker with the sound of his voice turned off, i.e. based on observation of the speaker’s facial expressions, posture and gestures and their changes in time.

And at the third stage, the assessments are already given according to the normal voiced video image of the speaker.

Upon completion of these three stages, the average values of each of the emotional-psychological properties and conditions of the speaker are calculated according to the data of all expert observers separately for the first, second and third stages of the speaker’s examination (and for clarity, the results are expressed as a percentage relative to the maximum assessment score) .

Then, the correspondence coefficient (K _resp ) of the auditory and visual scores given by expert observers to each of the speaker’s emotional and psychological properties and states of the speaker at the first and second stages of his examination is calculated by calculating the coefficient of rank correlation between the two series of these estimates Spearman (E. Sidorenoko. See the edition cited above).

The Spearman correlation coefficient calculated in this way is taken as the value of the correspondence coefficient (K _resp ) of auditory and visual assessments of sincerity-insincerity and other emotional and psychological properties and states of the speaker according to his non-verbal behavior and according to the main idea of the proposed method - is considered a psychophysiological measure (equivalent) sincerity and / or insincerity of the speaker, depending on the value of K _resp and its sign (since K _resp can take any value from plus 1 to min mustache 1, including zero values).

It is known that in mathematical statistics there is a certain criterion for assessing the significance of the correlation of two compared series of any quantities. The criterion of significance (we emphasize - in mathematics!) Conditionally accepted is a level of at least 0.05, which is calculated by special mathematical operations, and automatically in computer programs. Correlation coefficients having a significance level of more than 0.05 are considered not significant. This usually happens at medium and, as a rule, low values of correlation coefficients close to zero.

However, this formal rule, if we identify the degree of conformity of auditory and visual assessments with the Spearman coefficient, more precisely, with the equivalent matching coefficient (K _mat ), cannot be taken into account, since it does not relate to the essence of the task we have set. For we are not so much interested in the level of formal significance of the correspondence coefficient as its numerical value, telling us the degree to which correspondence is expressed, including the mismatch between auditory and visual assessments of the speaker’s non-verbal behavior, including all average values that are low and equal to zero (which indicates the complete discrepancy of the quantities of interest to us). Moreover, we are even interested in the negative values of K _respectively , which indicates a contradictory non-verbal behavior of the speaker (for example, the tone of the voice speaks of one emotional state, and facial expressions indicate another or nothing, etc.) and that is evaluated by experts- observers with high points of insincerity (see below).

Thus, in connection with the specifics of the task we have set, all the numerical values of the correspondence coefficients (K _resp ) are important to us; regardless of formal significance, both high, close to 1, indicating a high degree of conformity of the speaker’s auditory and visual assessments and a high level of his sincerity (as confirmed by expert observer ratings), and all medium and low K, _respectively , including zero and its negative meanings indicate different, lower levels of sincerity of the speaker or his insincerity.

The foregoing is illustrated by specific examples of assessing the matching coefficients of auditory and visual assessments of two speakers, presented in Figure 4 (speaker No. 9) and Figure 5 (speaker No. 4).

It can be seen that in speaker No. 9, as auditory assessments increase (the first stage of the examination), his emotional and psychological properties and conditions (dashed line 1), on the whole, visual assessments of the same qualities obtained in the second stage increase (solid line 2). This indicates a good agreement between the auditory and visual assessments of this speaker (i.e., the intonation of his voice corresponds to his facial expressions, posture and gestures), which is confirmed by his high positive coefficient of correspondence: K _corre = 0.92.

Speaker # 4 (Fig. 5) has a completely different picture: as his auditory grades increase (dashed line 3), visual assessments of his same emotional and psychological properties and conditions generally decrease (solid line 4). This speaks not only about the inconsistency of the intonation of his voice with his appearance, but also about the apparent contradiction between the auditory and visual assessments of his non-verbal behavior: the intonation of his voice says one thing, and facial expressions, posture and gestures - another (in terms of the severity of these qualities, with the exception of two qualities of "aggressiveness" and "calm", which he coincides in estimates). This contradictory picture is also reflected in the coefficient of correspondence, which this speaker (No. 4) has rather high, but negative value: minus (!) - 0.69.

Note: in Fig. 4 and Fig. 5, among the emotional-psychological properties and states of the speaker, there are the “disposing” and “non-disposing” properties (person). This was allowed in preliminary experiments, after a discussion of which the authors of the proposed method found it necessary to exclude these properties from the list of properties to be evaluated. At the same time, the list of properties recommended by the authors of the speaker’s states and conditions to be evaluated is replenished with a number of others: emotionality-indifference, slyness-honesty (see Figs. 1, 2, 3). One of the basic principles of forming a list of assessed properties and conditions of the speaker is the possibility of evaluating them both by voice intonation (first stage) and by the appearance of the speaker (second stage).

It is important to note that the significant differences found in the correspondence coefficients of speakers No. 9 and No. 4 according to the main idea of the proposed method indicate equally significant differences in sincerity, which is confirmed experimentally: according to general average assessments of expert observers: the sincerity of speaker No. 9 is 76 , 5%, and the sincerity of the speaker No. 4 is only 12.5%. Similar correspondences were found by us and other speakers.

Subsequently, on the basis of the average scores of sincerity given by expert observers at the third stage of the speaker’s examination, the total average value of scores of sincerity is calculated by summing the average scores of sincerity, sincerity, persuasiveness and honesty and dividing the result by the number of indicated close sincerity concepts (in this case on 4).

Similarly, the total average value of insincerity is calculated by averaging the average scores of insincerity, secrecy, unconvincing and deceitful.

This kind of averaging of close sincerity and insincerity of concepts (synonyms are selected from the dictionaries cited above by V. Dahl and S.I. Ozhegov) increases the accuracy of assessing sincerity and insincerity, since in a speech process the speaker’s sincerity and insincerity level and their severity, as a rule, quickly succeed each other, and the probability of evaluating sincerity-insincerity on just this semantically opposite pair may not be accurate enough for a given speaker.

Along with analogues of sincerity-insincerity, various possible emotional states are included in the number of properties and states of the subject to be evaluated, since sincerity is best recognized by the characteristics of the expression of emotions. Sincerely speaking, i.e. transmitting to the interlocutor his true thoughts and feelings (joy about something, sadness, anger, fear, etc.), according to the main idea of the proposed method, all elements of non-verbal behavior - voice intonation, facial expressions, gestures are activated quickly, often instantly and involuntarily “talk” about the same thing, i.e. well correspond to each other in the nature and degree of one or another experienced emotion. With an insincere speaker, i.e. imitating an emotion that he does not really experience or says that does not correspond to his thoughts, i.e. lying, such a harmonious, corresponding to each other complex of all elements of non-verbal behavior does not involuntarily arise, although he is trying to imitate (simulate) it. But insincerely speaking treacherously betrays, firstly, the unnaturalness of any elements of emotional behavior - intonations of voice, facial expressions, postures, gestures, and secondly, their inconsistency with each other. The contradiction between what a person thinks and feels and what he says manifests itself in the contradictory nature of his non-verbal behavior.

And finally, the final step in determining the speaker’s sincerity-insincerity is to multiply the total average value of the sincerity score by the correspondence coefficient (K _acc. ), And the insincerity is determined by multiplying its total average score by the expression (1-K _acc. ), Where K _acc is the coefficient compliance. For clarity, the final results of the assessment of the sincerity and insincerity of the speaker are expressed in%.

The following is a description of 11 speaker ratings.

Fig. 6 graphically shows the relationship between the correspondence coefficients (K _acc. , Line 5) for 11 speakers, average assessments of their sincerity excluding K _acc. (Line 6) and final general assessments of sincerity taking into account K _acc. (Line 7). It can be seen that a fairly good correlation takes place between these three lines: high values of K _correspond to high scores of sincerity of expert observers, and medium and low K _correspond to medium and low assessments of sincerity. This is experimental confirmation that the K value _accordingly determines a point estimate of sincerity not in the third stage of its examination.

On the same graph, one can see what kind of correction in assessing the sincerity of speaking by experts by observers gives K _resp : for all speakers, taking K _corre slightly reduces the overall assessment of sincerity, to a lesser extent for high levels of sincerity and more for low levels. In general, for this group of speakers (11 people), the correction for K, _{respectively,} ranges from 17% (for speaker No. 5) to 26% (for speaker No. 6) and 22% for speaker No. 4, which is very significant.

Figure 7 graphically shows the relationship between K _resp and average scores of insincerity. It is clearly seen, firstly: the smaller the importance of K, _{respectively,} for different speakers, the higher estimates of insincerity give them expert observers (lines 8 and 9). And secondly, the correction for K _acc. In this case increases the final assessment of insincerity by a value from 1.9% (for speaker No. 9) to 22.2% (for speaker No. 4, due to the large negative value of K _acc , indicating a large discrepancy between auditory and visual assessments by expert observers).

Thus, the results of testing the proposed method on a group of speakers (11 people) experimentally confirm its main idea: the dependence of the speaker’s sincerity-insincerity assessments on the degree of correspondence or inconsistency between different elements of his non-verbal behavior, i.e. perceived by the observer by ear or visually.

The psychological and practical meaning of the compliance coefficient is as follows. Assessment of expert observers at the third stage of the speaker’s survey, i.e. based on the perception of a normal voiced fragment of his video recording, it may be inaccurate, due to the extremely large amount of information about the sincerity or insincerity of the speaker (and the degree of severity of these properties and states) transmitted not only by the characteristics of the voice and motor forms of non-verbal behavior of the speaker, but also by their constant and rapidly changing relationships. Focusing attention, for example, on facial expressions, the observer can quite naturally miss some details of her discrepancy with gestures, posture or intonation of the subject's voice.

Therefore, the three times proposed by the authors assessment of the speaker’s sincerity-insincerity based on the same video fragment, reduces the flow of significant information at the first and second stages and, at the same time, allows you to more carefully evaluate the features of voice intonation (at the first stage) and expressive (expressive) speaker's movements (in the second stage).

The subsequent comparison of independent assessments of the same properties and states of the speaker, obtained at the first and second stages in the form of a coefficient of their correspondence (K _acc ), gives a generalized integral indicator of his sincerity-insincerity, calculated by a purely mathematical method (using Spearman's rank correlation coefficient) .

In connection with the foregoing, the compliance coefficient (K _acc. ), As it were, integrates in itself the results of the first and second stages, expressing these results with one number, i.e. the numerical value of K _respectively , which must be taken into account in the final assessment of the sincerity and insincerity of the speaker in the third stage.

For this reason, taking into account K _resp in the third stage of the speaker’s examination minimizes the likelihood of errors in evaluating his sincerity and insincerity by reducing proportionally to his numerical value (K _acc. ) Assessments of sincerity and increasing proportionally to the value (1-K _acc. ) Of the assessment of insincerity of expert observers.

Thus, the proposed method in order to improve accuracy differs from the prototype and analogues in a fundamentally new approach - combining the results of a three-fold assessment of sincerity and insincerity into one generalized indicator by applying To _resp .

Improving accuracy is also achieved:

1) by including in the number of evaluated properties and states of the speaker analogues of sincerity and insincerity;

2) the use of quantitative methods of assessment - psychological scaling and semantic differential;

3) by a special selection of expert observers.

The advantages of the proposed method also consist in a wider field of its application in comparison with the polygraph. Firstly, the ability to use it both in conjunction with a polygraph (in the form of comparing the results obtained), and as an independent method for assessing insincerity, i.e. "psychological lie detector."

Secondly, the possibility of quantifying the speaker’s sincerity as the most important advantage of speech, increasing its expressiveness and persuasiveness, which can be used in methods for improving public speech of politicians, public figures, TV announcers, managers, actors, vocalists and other specialists using speech as a "professional tool."

Claims (3)

1. A method for assessing the speaker’s sincerity-insincerity, including the video recording of the speaker, the assessment of his emotional and psychological properties and states during the speech by expert observers according to the characteristics of his non-verbal behavior, characterized in that the determination of the speaker’s sincerity-insincerity is made based on the results of a triple assessment of his emotional psychological properties and conditions for the same fragment of the video using psychological scaling, and at the first stage the assessment of emotional logical properties and states provide only the speaker at the off signal is a video image in the second stage on the basis of listening to voice - only the video image of the speaker switched off the sound of his voice, and in the third stage evaluation given by voiced video image of the speaker, and then calculates an appropriate coefficient (K _resp.) average scoring auditory and visual scores of expert observers given by them on the above emotional and psychological properties and states of the speaker on the first and second the tapas of his examination, by calculating the rank correlation coefficient between the two rows of these Spearman estimates, then the average values of sincerity and insincerity are calculated according to the estimates of all expert observers given by them at the third stage of the examination, and the sincerity is estimated by averaging the estimates of all psychological properties and conditions the speaker equivalent to the concept of “sincerity”, multiply the result by the coefficient of conformity and express the degree of sincerity of the speaker in percent by ratio to maximize its evaluation score and insincerity evaluation results produced by the averaging count insincerity and equivalent concepts, and thus obtained by the average value of the total insincerity multiplied by (1-K _{respectively.} ), where K _resp. - the coefficient of compliance and express the result as a percentage in relation to the maximum point of insincerity. 2. The method according to claim 1, characterized in that the group of expert observers is composed of at least 10 people with an emotional hearing of at least 80% according to the test of VP Morozov for emotional hearing. 3. The method according to claims 1 and 2, characterized in that among the emotional-psychological properties and states of the speaker, sincerity-insincerity, emotionality-indifference, frankness-secrecy, peacefulness-aggressiveness, calmness-anxiety, optimism-pessimism, persuasiveness are evaluated lameness, sensitivity-rigidity, frankness-craftiness, courage-fear.

Images