Testing Elias’s “Invisible Shame” Thesis
Thomas Scheff (4.6 k words + 5 graphs)
Abstract: The sociologist Norbert Elias’s The Civilizing Process examines European
history 14th to the 20th century: advice offered in English, German, and French etiquette
manuals (e.g., bedroom, bathroom, dinner). On this basis, he proposed that shame
replaced force as the main instrument of social control, but also became increasingly
unspeakable. This paper tests the latter conjecture 1800-2000 in five languages:
American and British English, French, German, and Spanish, using Google Ngrams. If
shame becomes less visible in modernization, then the term itself should occur less
frequently not only in speech, but also in writing. Although there are variations in slope
and shape in two graphs, the conjecture is supported in all five of the languages tested,
as shown in the attached charts. There is some turbulence at the beginning of the
Spanish chart, and much more all along in the German chart, but the downward slope is
fairly even in the first three charts, suggesting more stable societies. In modernization,
perhaps shame itself became shameful. It is possible that this can sometimes be the first
step in an endless series of recursive loops. If so, recursion of shame can lead to silence
and recursion of shame/anger to violence.
This paper tests a hypothesis about the visibility of terms for shame in five languages for two
centuries. In an extraordinary study (1939; 1978) of five hundred years of European history, the
sociologist Norbert Elias analyzed etiquette and education manuals in three countries. Two key
propositions resulted: 1. As physical punishment decreased, shame became increasingly
dominant as the main agent of social control. 2. As shame became more prevalent, it also
became unspeakable and virtually invisible because of taboo. There has been little response to
either proposition, but a new device, Ngrams, allows the second one to be tested systematically
with historical data.
The following excerpt gives the flavor of Elias’s study. It is from a nineteenth-century work (von
Raumer 1857) that advises mothers how to answer the sexual questions their daughters ask:
Children should be left for as long as possible in the belief that an angel brings babies.....
If girls should later ask how children come into the world, they should be told that the
good Lord gives the mother her child…"You do not need to know nor could you
understand how God gives children." It is the mother's task to occupy her daughters'
thoughts so incessantly with the good and beautiful that they are left no time to brood
on such matters.... A mother . . . ought only once to say seriously: "It would not be good
for you to know such a thing, and you should take care not to listen to anything said
about it." A truly well brought-up girl will from then on feel shame at hearing things of
this kind spoken of. (1978:180)
Elias first interprets the repression of sexuality in terms of hidden shame:
An aura of embarrassment…surrounds this sphere of life. Even among adults it is
referred to officially only with caution and circumlocutions. And with children,
particularly girls, such things are, as far as possible, not referred to at all. Von Raumer
gives no reason why one ought not to speak of it with children. He could have said it is
desirable to preserve the spiritual purity of girls for as long as possible. But even this
reason is only another expression of how far the gradual submergence of these impulses
in shame and embarrassment has advanced by this time. (1978:180)
Elias raises a host of significant questions about this excerpt, concerning its motivation and its
effects. His analysis goes to what may be a key causal chain in modern civilization: denial of
shame and of the threatened social bonds that both cause and reflect that denial.
Considered rationally, the problem confronting him [von Raumer] seems unsolved, and
what he says appears contradictory. He does not explain how and when the young girl
should be made to understand what is happening and will happen to her. The primary
concern is the necessity of instilling "modesty" (i.e., feelings of shame, fear,
embarrassment, and guilt) or, more precisely, behavior conforming to the social
standard. And one feels how infinitely difficult it is for the educator himself to overcome
the resistance of the shame and embarrassment which surround this sphere for him.
(1978:181)
Elias's study suggests a way of understanding the social transmission of the taboo on shame.
The adult teacher, von Raumer, in this case, is not only ashamed of sex, he is ashamed of being
ashamed. The nineteenth-century reader, in turn, probably reacted in a similar way: being
ashamed, and being ashamed of being ashamed, and being ashamed of causing further shame
in the daughter. Von Raumer's advice was part of a social system in which attempts at civilized
delicacy resulted and continue to result in an endless chain reaction of hidden shame.
Elias understood the significance of the denial of shame to mean that shame goes
underground, leading to behavior that is outside of awareness:
Neither rational motives nor practical reasons determine this attitude, but rather the
shame (scham) of adults themselves, which has become compulsive. It is the social
prohibitions and resistances within themselves…that makes them keep silent.
(1978:181; emphasis added)
Like many other passages, this one implies not only to a taboo on shame, but the actual
mechanisms by which it is transmitted and maintained.
How can shame increasingly permeate social life, yet at the same become less visible? In the
second passage above, Elias referred to one way, the use of what he called circumlocution.
Instead of using the s-word, one may say “I feel hurt, worthless, rejected” or “like a nobody.” In
her study of emotions in psychotherapy, the psychologist Helen B. Lewis (1971) named these
usages “overt, undifferentiated shame.” This usage hides shame, but with only a thin veil,
because the terms don’t hide the pain completely.
Lewis goes on to describe another way to hide feeling that she calls “bypassed shame.” In these
instances, pain is completely hidden. One way involves hiding shame behind anger or anger.
Another way is to withdraw into silence. If a person is insulted, they may indicate humiliation
indirectly by saying “You hurt my feelings.” But they are more likely to not respond at all, or
answer in kind: “No, it’s not my fault, it’s yours, you sonofabitch!” Lewis’s work will be
discussed further below. Elias and Lewis both make the point that brazenness particularly, the
seeming absence of shame in modern societies, is an illusion.
Historical Decline the Use of Shame Terms
If, as Elias proposed, shame became increasingly unspeakable in modernization, this change
should be found not only in spoken, but also in written language. The five charts in the
Appendix were obtained from Google Ngrams, based on millions of digitalized books (Michael,
et al 2011). All of the graphs support the Elias conjecture, since they all show decline in the use
of the term for shame in the five languages. The decline of the use of the s-word is remarkably
even over the course of 200 years, and similar, according to the Ngrams, in three of the five
languages: American English, British, and French.
The Spanish and German cases are similar in that both show decline in usage, but are
substantially different in two ways. First, usage is on a much smaller scale than in the English,
British and French cases. In the beginning of the period the highest frequency for Spanish is
only about a fifth (approximately .001) of those in the first three graphs (the average is about
.005). The earliest high in the German case is still lower, about .00001, only a tenth of the
American, British, and French average.
There is an even greater difference in the lowest frequencies at the end of the period. Spanish
is again only about one fifth (.0002) of the average of the first three graphs (a little more than
.001), and German is still less, about a hundredth (00001). At this point these large differences
of scale between the two groups are puzzling. One possibility is that if the number of books
used is considerable less than for the English and French cases, the graphs for the Spanish and
German would be less reliable than those for the the first three countries.
The second difference between the two groups is in the shape of the graphs. Those for the first
group are relatively stable, but the Spanish graph is somewhat erratic at first, the German much
more erratic, and over the entire period. The rapid shifts might also support the idea that the
Spanish and German data is smaller than the other group, and therefore less reliable
This difference is confirmed for the 1800-1899 period by Chart S4 in the Michel et al article
(2011). The number of words is much smaller for the Spanish and German graphs than for the
other three. However, there is little difference between them for the period 1900-1999.
It seems more likely that the difference of the two graphs is mostly historical. Perhaps much of
the sudden change in the Spanish graph 1800-1840 was caused by the loss of its colonies, and
the differences that emerged between languages in Spain and its former Latin-American
colonies.
The German Changes
The rapid changes in shape are even greater in the German case, and occur in the whole graph,
not just in the beginning. Although it also shows a decline, the term schande has undergone
many sudden changes in the two hundred year period. From my earlier study that involved
German history (1994) I can interpret one of the changes, the rapid downturn in 1945.
Schande (intense shame) was one of Hitler’s favorite terms. For example, he usually referred to
the democratic German government he displaced, the Weimar Republic, as "Fourteen years of
shame and disgrace." In German this phrase carries much greater force than the English
translation: Vierzehn Jahren von Schmach und Schande! It almost requires shouting to get it
out, which is exactly what Hitler did. He used it as an epithet. But when Hitler was removed
from power, his pet terms lost their charm.
For commentary on the whole pattern of the German graph, I called upon my colleague, Ursula
Mahlendorf, who is an expert on German history and language.
In contrast to the other four, Germany was not one country until 1872, until then the
states were independent political, social and economic actors, much to the detriment of
commerce and trade, causing interstate hostility and social unrest. Even the 1848
revolution (imitating the French revolutions of the 1830-40’s), though democratic, was
different in each state.
At the end of these failed revolutions was the Frankfurt- Paulskirche constitutional
convention that was to give all of Germany with the exception of Austria a, democratic
constitution and central government. The enterprise failed, the regional rulers returned
and held power till the Prussians, the largest of the states, forced the union during three
wars in the 1860s and 70s, the war with Denmark for the Northern border; the war with
Austria for the Southern border and Prussian supremacy; and the Franco-Prussian war
when a German coalition of princes defeated France under Prussian leadership; then
Prussia simply proclaimed the Prussian king emperor and the other states acceded to
become parts of the Reich. So it is only from then on that there was a national economic
and social policy. That also meant that there was no substantial industrialization till the
1870s; which then took off rapidly at great social expense.
Major disruptions of the social and political order in the early part of the century
occurred because of famines (in the 1800-1810s, the 1830s; 1845 to 1850. Wars:
Napoleonic 1797-1806; 1812-15; the wars of the 1860s; the Franco-Prussian; the first
WW.
Economic down-turns with loads of bankruptcies occurred in 1873; 1894;1905; 1919;
1923; 1929-33; 1945; in 1948 the devaluation of the currency 10:1 made possible the
rebuilding of the economy. By 1949 the West German economy was helped
substantially by the Marshal plan and from then on it conformed to that of the West
European states (with the exception of Great Britain). The East German economy after
being nationalized and denuded by steep reparations to the Soviet Union had a brief upturn during the 1950s and then limped along and gradually failed until it was bankrupt in
1990.
These remarks suggest that the great difference of the shape of the German graph might be
explained in the volatile nature of the society it represents, in contrast with the more or less
stability of the other four countries.
A Further Conjecture: Recursion of Shame
It is conceivable that that the decrease in the use of the shame term is caused by a taboo. The
psychologist Gershen Kaufman is one of several writers who have argued that shame is taboo in our
society:
American society is a shame-based culture, but …shame remains hidden. Since there is shame
about shame, it remains under taboo. ….The taboo on shame is so strict …that we behave as
if shame does not exist (Kaufman 1989).
Kaufman’s phrase, shame about shame, may have meaning beyond what he intended. As fear
sometimes leads to more fear, causing panic, shame about shame can loop back on itself to
various degrees, even to the point of having no natural limit. Recursion of shame will be
discussed further below. Suppose that shame is usually hidden, as suggested by the idea of taboo.
What difference would it make?
A taboo could lead to a general weakening of the whole society, to the extent that shame about
shame becomes recursive, leading to endless loops. My own interest in recursive emotion
began long ago in teaching university classes on emotions. When we discussed embarrassment
and blushing, there were usually one or two students who complained that their blushing made
them miserable. When they became aware that they were blushing, they would be further
embarrassed, no matter the cause of the first blush.
This comment by a 20-year-old female student provides an example:
I often blush when I receive a compliment. On one occasion a friend praised my smile. I
immediately felt a blush. Then my friend said “Oh, you are blushing!” I said “Yes, I can
feel it!” On some occasions my blush feels as if it will be eternal.
With these kinds of observations as background, I was struck by a story told by the noted actor
Ian Holm. On one occasion he had muffed his lines, but when he became aware that he was
blushing, he blushed more. The more he became embarrassed by his blushing, the more he
blushed and the more embarrassed. This process went on until he ended paralyzed in the fetal
position, requiring that he be carried off the stage.
Holm’s story points to an emotion process that is internal, implying that emotion loops might
have no natural limit. This idea is also suggested by the student’s comment above, when she
states that her blushes sometime feel eternal. Another process, social alienation, may also be
recursive, but an external process. Audience members in a theatre fire could become afraid
because they are afraid themselves, and they see other audience members afraid, resulting in
loops within and between persons causing more fear. Social-emotional responses leading to
more social and emotional reactions may result in chain reactions.
Recursion: Isolation Traps and Feeling Traps
In this section, two main kinds of recursive loops will be considered: a social loop of
rejection/isolation on the one hand, and a shame loop (Lewis 1971), on the other. The idea of a
rejection/isolation loop is straightforward. Being or even just feeling rejected leads toward
alienation, and the more alienated, the more likely further rejection, a spiral. This process is
mostly social rather than psychological.
The part played by emotion loops is more subtle. It seems to be based on shame, but the kind
of shame that goes unnoticed. Lewis noted that when shame occurs but is not acknowledged, it
can lead to an intense response, a feeling trap: one becomes ashamed of one’s feelings in a way
that leads to further emotion. Since normal emotions are brief, Lewis’s idea of a feeling trap
opens up a new area of exploration.
The particular trap that Lewis described in detail involved shame/anger sequences. One can
rapidly become angry when ashamed, and ashamed that one is angry. She called the result
"humiliated fury.” In her study of psychotherapy sessions, she found many word-by-word
instances of episodes in which unacknowledged shame was followed by either hostility toward
the therapist or withdrawal. In her examples of the latter, withdrawal takes the form of
depression. (1971, p. 431 and passim). In a later volume ( 1987, pp. 29-49), Lewis reviewed
many studies by other authors that showed strong correlations between shame and
depression.
How does shame become destructive? It may be that when shame is kept secret,
(unacknowledged, in Lewis’s language), there is little chance that it will be resolved. It is
possible that normal shame is resolved through verbal means, the telling or thinking of the
incident, often repeatedly, and through humor. However, in the case of intense humiliation,
lengthy verbal or at least conscious consideration might be needed before any humor can be
found in the offending incident.
In many of my classes, I have asked student volunteers to tell the class about their most
embarrassing moment. Invariably some of the volunteers, during the course of their story,
become convulsed with laughter. Perhaps good natured laughter signals the complete
resolution of unacknowledged shame.
It appears that when shame is not resolved, it can build up a backlog. When there is
considerable backlog, then any new incident is felt in itself, but also seems to reactivate the
backlog, making the new incident, even if seemingly trivial, extremely painful. Backlogs of
shame can lead to trouble.
Emotion Spirals
Lewis’s idea of humiliated fury as a feeling trap can be a first step toward a theory of the
emotional origins of either withdrawal or aggression. Since none of the therapy sessions she
studied contained physical aggression, she didn’t consider the kind of feeling traps that could
result in violence, on the one hand, or long lasting or total withdrawal (as in clinical depression).
Lewis described feeling traps as emotion sequences. These sequences involve at most three
steps, as in the case of the shame/anger sequence to depression. It will be necessary, however,
to go beyond three steps, even as far as an endless spiral. Such a process would be a doomsday
machine of interpersonal and inter-group withdrawal or violence. The combination of isolation
and denial of shame can lead to self-perpetuating loops that generate either complete
withdrawal or violence.
Some emotion sequences may be recursive to the point that there is no natural limit to their
length and intensity. As already indicated, blushers provide an everyday example. This feeling
trap would not be a shame/anger spiral, but rather shame/shame: being ashamed that you are
ashamed, etc.
Recursive shame-based sequences, whether shame about anger, shame/fear, or shame about
shame need not stop after a few steps. They can spiral to the extent that they rule out all other
considerations. Collective panics such as those that take place under the threat of fire might be
caused by shame/fear spirals, one’s own fear and that of others reflecting back and forth can
cause still more fear, leading to a triple spiral: a spiral within each person, and a spiral between
them.
Although Lewis didn’t consider the possibility, depression might be a result not only of a
shame/anger spiral, but also shame/shame, or a combination of both. Judging from her own
transcriptions, withdrawal after unacknowledged shame seems to be much more frequent than
hostility toward the therapist.
The less frequent shame/anger spiral, humiliated fury, or a shame/shame spiral with the anger
hidden, might be basic causes of violence to the extent that they result in self- perpetuating
loops. A person or group caught up both in alienation and in a shame-based spiral might
become oblivious to all else, whether moral imperatives or danger to self or to one’s group.
Limitless quarrels or withdrawal can be generated by shame/anger and/or shame/shame spirals
within each party, and an isolation spiral between them.
It is conceivable that shame spirals could be a predominant cause of violence, with
shame/anger playing only a hidden part. This might be the case in killings that are carefully and
lengthily premeditated as well as instantaneous violence. Shame spirals and shame/anger
spirals could be equally involved.
Recursion of Emotions and Alienation in Killers
It has been suggested that recursive thinking is unique to human beings, differentiating their
mental processes decisively from other species (Corballis 2007; 2012). The theory presented
here proposes that recursion of feelings, feeling about feeling, would also differentiate humans
from other species, and explain episodes of depression or rage, silence or violence. As already
indicated, recursion can also occur in alienation, leading to complete isolation or engulfment.
The idea of two types of alienation in connection with killings done by two persons jointly must
be considered here. The founder of family systems theory, Murray Bowen (1978), distinguished
between two kinds of dysfunctional relationships, engulfment (fusion) when the bond is too
tight, and isolation, when it is too loose. In engulfed relationships, one or both parties
subordinate their own thoughts and feelings to those of the other(s). In true solidarity, each
party recognizes the sovereignty of the other, but balances respect for the other’s position with
respect for one’s own, no more and no less.
Elias's (Introduction, l987, pp. 77-79,) discussion of the "I-self" (isolation), the "we-self"
(engulfment) and the "I-we balance" (true solidarity) makes the same point. Elias proposed a
three-part typology: independence (too much distance), interdependence (a balance between
self and other), and dependence (too little distance).
Engulfed relationships are alienated because at least one of the parties gives up vital parts of
the self in order to be loyal to the other party. That is, one or both parties are alienated from
self, in service to the other. In this kind of relationship, the kinds of negotiations that can be
called upon by two independent parties are lost. All of the pairs of killers that I have examined
seem to have been alienated in this mode. One person dominates the other. This was certainly
the case of Eric Harris and Dylan Klebold in Columbine. Eric completely dominated Dylan, and
Dylan idolized Eric (Larkin 2007, pp. 144-148).
For the theory to be generally applicable, a further problem needs to be engaged. Why do
certain people and groups end up in cybernetic loops of alienation and shame, but not others?
Most individuals and groups in modern societies probably fail to fully acknowledge much of
their alienation and shame/embarrassment/humiliation. The key may lie in the issue of fullness
of acknowledgment.
Lewis (1971) and other shame researchers have considered acknowledgment only as a
dichotomy, yes or no. Perhaps even even a slight degree of acknowledgement avoids
continuous recursion. Persons and groups that manage to stamp out any acknowledgment of
shame and alienation would then be on the path of endless recursion, and therefore to
withdrawal or violence
There are several studies that suggest that shame/anger, even when the anger component is
not obvious, can be so painful and controlling as to lead to murder and suicide. The clearest
examples are Websdale’s (2010) cases of familicide (the killing of one’s spouse and one or more
of the children) in the US over the last 50 years. The study shows that most of the killers
seemed driven by secret shame. The study is large, detailed, and systematic, showing evidence
of intense shame in almost all cases. This type of murder is a multiple killing, but usually
enclosed within a single family. These findings strongly support Gilligan’s (1997; 2011) idea that
violence is caused by shame.
The theory outlined here proposes shame as one causal agent and alienation as the other.
When these two components interact without limit, the stage is set for either withdrawal or
violence. Fortunately for the survival of the human race, withdrawal seems to be by far the
most frequent reaction.
Anticipation of loss of control and/or unbearable pain might lead people to avoid emotions
entirely. This kind of avoidance also may have still another kind of looping effect: emotional
backlogs. The more avoidance, the more the bodily buildup of emotional tension. The more
backup, the greater the pain that is anticipated, which can lead to an avoidance loop.
Conclusion
This paper has tested the hypothesis that the term for shame in five Western languages has
become less and less visible in written language in the course of modernization. The Ngram
charts support the hypothesis for the languages tested: in all five, the frequency of the terms
for shame decreases. There is also a preliminary attempt to explain why the shapes of the
graphs of American English, British, and French shame terms are so similar, and why those for
Spanish and German are different.
Since the study found independent confirmation that shame has become less visible in
modernization, a further hypothesis is proposed to account for the findings. Shame becomes
invisible because of recursion: as people become ashamed of being ashamed, the process may
continue to the point that shame becomes completely hidden through recursive loops, ending
in silence or violence. This cybernetic idea is supported by Retzinger’s systematic study (1991)
of the moment by moment exchange of emotions in marital quarrels. The idea should be
further tested for individuals and groups, since it could be central to understanding both
individual and societal withdrawal and conflict in our current civilization.
References
Corballis, Michael. 2007. The uniqueness of human recursive thinking. American Scientist 95,
May-June, 240-248.
_______________2012. The Recursive Mind: The origins of human language, thought, and
civilization. Princeton N. J.: Princeton University Press
Elias, Norbert. 1939. Über den Prozess der Zivilisation). Reprinted in 1978 as The Civilizing
Process. London: Blackwell.
Elias, Norbert. 1987. Involvement and Detachment. Oxford: Blackwell.
Gilligan, James. 1997. Violence – reflections on a national epidemic. New York: Vintage Books
_____________ 2011. Some Politicians are More Dangerous than Others. New York: Polity
Gottschalk, Louis, C. Winget. and G. Gleser. 1969. Manual of Instruction for Using the
Gottschalk-Gleser Content Analysis Scales. Berkeley: UC Press.
Kaufman, Gershon. 1989. The Psychology of Shame. New York: Springer.
Larkin, Ralph W. 2007. Comprehending Columbine. Philadelphia: Temple University Press.
Lewis, Helen B. 1971. Shame and Guilt in Neurosis. New York: International Universities Press.
_____________1987. The Role of Shame in Symptom Formation. Hillsdale, N.J.: LEA.
Jean-Baptiste Michel, Yuan Kui Shen, Aviva Presser Aiden, Adrian Veres, Matthew K. Gray, The
Google Books Team, Joseph P. Pickett, Dale Hoiberg, Dan Clancy, Peter Norvig, Jon
Orwant, Steven Pinker, Martin A. Nowak, and Erez Lieberman Aiden. 2011. Quantitative
Analysis of Culture Using Millions of Digitized Books. Science 331 Supplemental
Materials.
Retzinger, S. M. 1991. Violent Emotions: Shame and Rage in Marital Quarrels. Newbury Park:
Sage.
Scheff, Thomas. 1994. Bloody Revenge: Emotions, Nationalism and War. Boulder: Westview.
_____________2012. A Social/Emotional Theory of “Mental Illness.” Intern’l Journal Social
Psychiatry. June 12. 37-53,
Websdale, Neil. Familicidal Hearts: The Emotional Style of 211 Killers. Oxford: Oxford University
Press.
Von Raumer, Wilhelm. 1857. The Education of Girls. (Cited in Elias, l978).
Appendix
American English
British English
French
Spanish
German