1
RESEARCH REPORT ON DST PROJECT - 2015
Rukmini Bhaya Nair and the 'Language, Emotion, Culture' Group, 2010-2014
Indian Institute of Technology Delhi, IITD*
CONTENTS
PAGES
I. Summary of Research Questions, Some Hypotheses and Project Goals
1-4
II. Empirical Research Plan and Description of Data
4-6
III. PAPER Beta Version, including Tables and Graphs
7-53
1V. Significant Findings (part of paper above)
47-52
V. Coda and Acknowledgements (part of paper)
52-53
VI. Bibliography (More than 4000 refs – not attached but please request if required)
VII. Appendices 1-4 (Not attached but available on request)
I. SUMMARY
This work presents some empirical findings from a three-to-four-year research project located
in India that seeks to approach the complex notion of ‘cognition’ by studying two key areas
of cognitive processing – namely, language and emotion.
---------------------------------------------------------------------------------------------------------------* Our group was funded by the Cognitive Science Initiative (CSI), Department of Science and Technology
(DST), Govt. of India. The duration of this particular project was
March 2010 to mid-2013, later
extended to March 2014. The Project Co-PIs were: Professor Purnima Singh, Dr. C.A. Tomy, both
faculty at IIT, Delhi. Project Associates and Assistants, past and present, who have greatly contributed to
the project include Dr. Souvik Mukherji, Dr. Ashraf Bhat, Dr. Renuka Jain, Nivida Chandra, Swathi
Yemmanur, Yukti Kumar, Sakshi Oberoi, and Abhishek Upadhyay. We also gratefully acknowledge the
cooperation of Dr. Falk Huettig of the Max Planck Institute (MPI), Nijmegen. Our advisory board
included V S Ramachandran (San Diego) and Steven Pinker (Harvard) and several others who were very
helpful in forwarding the project – see Acknowledgements for a complete list.
2
Our point of departure for this research was the increasingly controversial hypothesis that the
mind/brain is highly modular with specialized systems for each cognitive function (Chomsky,
1990, 2002; Fodor, 1985; Coltheart, 1999). Thus, our initial question was: Could there be an
Emotion Acquisition Device (EAD) similar to the Language Acquisition Device (LAD)
that has long been postulated as the human brain's most distinctive cognitive system? In more
detail, the broad questions we considered included the following:
a. Can we infer the existence of an ‘emotion acquisition device’ at par with Chomsky’s
‘language acquisition device’ through a detailed study of the verbal and non-verbal behaviour
of the children whose emotional behaviour our databases are designed to capture?
b. If there is indeed something like an emotion 'module', how does this putative module
developmentally interface with the language module? What is the nature of this interaction?
c. What are the milestones of emotional acquisition in children? Is there a 'critical period' for
emotional development which makes later, more agentive, emotional actions possible?
d. Can we postulate a set of basic structures for the 'emotional act' similar to those suggested
for the 'speech act' (classes of illocutionary act, locutions). If so, are Austinian 'perlocutions'
the best candidate points at which we might conjoin illocutionary acts and emotional actions?
e. In what sense do narratives, which belong to FLB (the faculty of language broadly defined,
Hauser, Chomsky and Fitch, 2004) share the recursive properties of FLN (the faculty of
language narrowly defined)? If narratives are in fact 'recursive' in their structure are they the
fundamental discourse structures that produce, drive and explicate emotional knowledge?
f. How can ideas of 'cultural survival' be explored through narrative and notions of stigma?
To elaborate, with respect to e. and f. above, one of the places to look for adaptation and the
development of adaptive traits is within culture, where the human species is reflexively
and rapidly effecting changes in its own environment (see Deacon, 2007, 2011). Our
objective in this ‘extensions and cultural implications' study were to:
 try and understand how humans use the putatively interacting modules of
language and emotion to survive the often harsh cultural challenges posed by the
normative rules that every society imposes on its members.
3
 identify and analyze the cognitive strategies that human actors use to ensure
cultural survival in situation of cultural ‘threat’ or danger that are inevitably also
emotionally charged
 record how narratives as linguistic constructs are used by humans for the
purposes of cultural navigation and cultural survival, given that language is
held to be the distinguishing feature of homo sapiens sapiens and our hypothesis
(Nair 2002, 2003, 2009, 2011) is that narratives are recursively designed at the
level of discourse (FLB) in the same way as sentences (held to be the
distinguishing aspect of the language ‘device’ by Chomsky and other generative
grammarians) are recursive at the level of grammar (FLN)
1b. SOME TENTATIVE HYPOTHESES
 EMOTION ACQUISITION HIERARCHY: Given our assumption that language
and emotion are slow-growing across time, there will be parallels with language
development in that there will be stages in emotional development with the ‘basic
emotions’ (happiness, sadness, surprise, fear, anger, disgust) followed later by more
complex emotions (such as shame and guilt).
 EMOTION RECOGNITION: Given our assumption that there are a range of
emotions in the cultural sphere, mothers will be typically be able to accurately
identify these emotions not only in their own children but in others.
 EMOTIONAL RECALL: Given our assumption that mothers are ‘best observers’ of
their children, they will be able to recall the stages at which their children acquired
various emotions.
 EMOTIONAL RECORD AND NARRATIVE EMBEDDING: Given our
assumption that the language ‘module’ is ‘intelligent’ at least as far as the FLB is
concerned, we expect that mothers will use narrative and other expressive discourse
strategies to recall their children’s emotional expressions
1c. AIMS OF PROJECT
Original Statement of all objectives over the 3-4 Year Duration of this Project.
4
1. Generating a database for emotional studies based on mother child interactions,
including qualitative narrative data. OBJECTIVE ACHIEVED.
2. Production of a documentary research archive of 16 ethnographic films on the
spontaneous cultural expression of each of Darwin’s sixteen listed emotions.
OBJECTIVE ACHIEVED (WE HAVE ACTUALLY PRODUCED 20 FILMS)
3. Description of some continuities and discontinuities in emotional expressions in
culture. OBJECTIVE ACHIEVED. WE HAVE RESULTS ON 24 EMOTIONS.
4. Conducting a variety of social psychological laboratory experiments to determine
how devalued (stigmatized) social identities present ‘cultural survival’ threats and
how people cope with them. WE HAVE NOT YET FOLLOWED UP ON THIS
OBJECTIVE.
5. Correlation of the self-representation of the body in the Penfield motor-neuron
strip with linguistic expressions of emotion. WE HOPE TO CONDUCT THESE
AND OTHER EXPERIMENTS OVER THE NEXT YEARS - SUGGESTIONS
VERY WELCOME!
II. OUR EMPIRICAL RESEARCH PLAN WAS:
 to consider the putative co-development of language and emotion in children by
involving children’s mothers as 24X7 ‘best observers’
 to undertake a large cross-sectional survey of 500 mothers with children in three
evenly divided age groups (0-30 months, 31-60 months, 61-100 months)
 to prepare an ‘emotion-language’ questionnaire and interview schedule consisting
of close-ended questions as well as open-ended interviews that would be
amenable to both quantitative and qualitative analyses (using, for example, SPSS
& MAX QDA software)
 to create a robust database of how mothers recognize, recall and describe their
children’s emotions across an emotional spectrum of 24 emotions.
IIa. FIELD WORK:
Ours has now turned out to be one of the largest studies of emotional and language
development in the world. The results from it seem to confirm a clear developmental order
5
in the acquisition of an emotional repertoire as well as some statistically significant
findings on the influence of culture on emotional cognition.
Our main corpus consisted of hour-long audio-recordings as well as answers to detailed
questionnaires (500-600 hours total). Our respondents were 505 Hindi-speaking mothers of
children aged 3 months to 8 years. Further, the emotions we considered, unlike in most
previous studies, were not just restricted to the basic emotions but totalled 24 in all.
The research method we followed in this cross-sectional study is described in detail in the
paper below. Indeed, the methodology we chose to adopt of some import as it is at the heart
of a current critique of methodology in psychology and the cognitive sciences which
routinely privilege 'objective' measures above introspection (Billig, 1998; Cohn, 2008, 2010;
Hatfield, 2005; Langlitz, 2010; Martin, 2010; Prinz, 2004; Young, 2012). Since this
preference is especially ironic in the area of affective science, our research has consciously
sought to ameliorate this bias. This segment of our research has therefore positioned not so
much the researcher but the mother as 'best observer' in all our 505 mother-child dyads.
Mothers' recall, recognition and narrative recounts of the expression of emotions in
their own children was seen by us as a fundamental tool to understand emotional
development, based on the intuition that as a general rule of thumb mothers are (have to be?)
deeply intimate with their children and recognize in both gross and finely nuanced ways
the embodiment of emotions in them in order to attend to a simple but not necessarily
simple to achieve, goal: the survival and well-being of her child. Further, since our corpus
was large, the results we've got seem statistically reliable.
Thus, if the claim made in a relatively recent Harvard working paper about 'children’s
emotional development [being] built into the architecture of their brains' (Working
Paper No. 2 http://www.developingchild. net © 2004, National Scientific Council on the
Developing Child, Center on the Developing Child, Harvard University) is to be taken
seriously, then the current research could throw some new light on this well-worn
'innateness hypothesis.' We expand on these preliminary remarks in the paper below (pages
7-40).
IIb. ETHNOGRAPHIC FILMS: A DARWINIAN ENTERPRISE?
6
Unlike other faculties (touch, taste, vision, hearing and smell) which are more or less fully
functional within one year or so of birth, language is relatively slow-growing, with the full
range of language abilities acquired only by about 4 years of age or so.
Our general hypothesis in this research was that emotions, too, develop relatively slowly so
that children acquire their cultural grammars of emotion in tandem with their language
grammars within specific 'contexts of use'.
Here, we acknowledge the seminal influence of Charles Darwin's The Expressions of
Emotions in Man and the Animals (1872) which kick-started, in a sense, the entire field of
modern cross-cultural emotion research. It will be recalled that Darwin sent his 'emotion
questionnaire' (see Appendix I) to respondents in all five continents, encompassing over
thirty countries, since he sought to find evidence for his evolutionary hypothesis that certain
'basic' emotions are shared across cultures (not to mention species!). In our follow-up on this
'Darwinian' strain of enquiry, we have initiated two sub-projects:
IIbi. Ethnographic films:
We have made a set of 20 ethnographical documentaries of about 15-25 minutes in
duration in Indian families which try and capture spontaneous expressions of, and 'folk' views
on, the basic as well as non-basic emotions (happiness, disgust, surprise, shame, reflection
etc.) so that we can study the micro-expressions associated with the 'rise-and-fall' of
emotional episodes and actions as well as analyse the narratives of emotions expressed in
everyday life. This is certainly the largest world repertoire of such films in a non-western
context and is a database that we are sure will be of interest to researcher in other cultures.
IIbii. Darwin's questionnaire redistributed:
We have begun to redistribute Darwin's original 1872 questionnaire to dancers trained in
classical Indian dance and actors as well as a 'control' population of non-dancers and nonactors with interestingly different results from the ones reported by Darwin (see Appendix II).
IIc. CROSS-CULTURAL DATA:
We recognize, of course, that nothing necessarily hinges on the 'largeness' of databases
but that much depends on the manner in which we 'capture' emotional salience. In
pursuit of this objective, we have collected and analysed data from a large scale (100
responses each from 40 Dutch-speaking subjects in Nijmegen, Holland and 40 Hindi
speakers in New Delhi, India = 8000 responses). This study, conducted in collaboration with
Dr. Falk Huettig of the Max Planck Institute, has provided an important insight into a crosscultural understanding of emotion and its linguistic expression. We find that Hindi and
7
Dutch speakers varied widely and systematically in this study when we asked them to
associate emotion versus non-emotion abstract nouns with the first verb that came to their
minds. Some results from this experiment are mentioned in our 'paper' below.
III. PAPER – BETA VERSION
WHAT'S BASIC ABOUT BASIC EMOTIONS? A RESEARCH BASED RE-EVALUTION
Introduction: In an important discussion paper written over two decades ago, Ortony and
Turner (1990) suggested that there are no principled reasons why certain emotions are
treated as 'basic'. Researchers, they contended, had varied widely both in identifying basic
emotions and in enumerating them, postulating just 2 (pleasure and pain) in some cases to
as many as 18 in others. This question-mark over the status of the 'basic emotions' has
persisted. However, researchers over the years (see Levenson, 2011, for an efficient review)
seem to have settled on six or seven basic emotions, namely: happiness, surprise, sadness,
fear, anger and disgust; and a few have suggested that love or 'trust' should be accounted a
basic emotion as well. In the current paper, we present our own research on the basic
emotions and their developmental order in children using a methodology that we believe is
innovative as well as particularly suited to a study of emotional acquisition. Before we
describe our results, however, we explain our method, beginning with why we chose the 24
emotions that we did.
METHODOLOGICAL CONSIDERATIONS:
i. Choice of Emotions for Investigation: First, we included, for purposes of consistency
and comparison all sixteen 'emotions' in Darwin's original questionnaire (see Appendix II)
including a film on ways of saying 'yes' and 'no' across cultures, which some would contend
is hardly indicative of emotion on the face of it (no pun intended). The emotions Darwin
asked respondents about were: 1. astonishment or surprise; 2. indignation; 3. shame; 4.
reflection; 5. 'good spirits' or happiness; 6. 'low spirits' or sadness; 7. sneering; 8. contempt; 9.
obstinacy; 10. disgust; 11. fear; 12. laughter; 13. helplessness or inability; 14. sulkiness; 15.
guilt/jealousy; 16. yes/ no.
It is, however, apparent from Darwin's list that it has overlaps and makes some unusual, even
idiosyncratic, choices. In such cases, we exercised our judgment. For example, 7. and 8.
8
'sneering' and 'contempt' are semantically close, as are 9. and 14. 'obstinacy' and 'sulkiness'. In
these cases, we ran these 'emotions' together and only used one term for these in our study.
To illustrate further: While we had, for example, a commonly used term in Hindi/Urdu for
obstinacy ('zidd'), we felt that 'sulkiness' was quite an 'English' expression with a specific
cultural history which might confound Indian mothers.
Also, the question arose at a more conceptual level as to whether 'obstinacy' and 'sulkiness'
were in themselves emotions in the way 'fear' or 'anger' intuitively are or whether they were
'expressions' of mental states not amounting to emotions. Sulkiness in this respect might
constitute a behavioural response to a situation where a child was not getting what he or she
wanted rather than an ‘emotion proper’. Worth noting, too, is the fact that sulkiness seems
conceptually ‘gendered’ in that it is hardly ever associated with an adult male. We felt that
these questions about semantic overlap and cultural context could be answered in three
ways:
One, we could approach the question theoretically and argue that the expressions of emotions
could not be separated from the emotions themselves as William James and, later, Carl Lange
did (see Appendix III for a chart of various theories of emotions from the early 20th century
to the present). In this case, we would really be arguing, for example, that if obstinacy and/or
sulking showed distinctive behavioural characteristics, then they were in fact different
emotions. Two, we could independently confirm our semantic intuitions that expressions of
a particular emotion overlapped with another by questioning cultural insiders. In the case of
'zidd' or obstinacy, for instance, we asked our informants (native speakers of Hindi) whether
they thought 'zidd' included sulking (here, our informants emphatically held that the powerful
cultural construct of 'zidd' could easily include 'sulking'). Three, we could proceed by
looking at previous research and ascertaining whether other researchers has focussed on these
emotions (as it happens, we drew a blank here in the case of Hindi/Urdu 'zidd').
With these concerns in mind, we therefore augmented Darwin's original fifteen emotions with
some others. These included some culturally inflected emotions listed by Bharata in the
Natyashastra, an ancient but still key text on the emotions in the Indian context; for example,
pride, courage and peacefulness, which were not emotions mentioned by Darwin in his
questionnaire. In addition, we included certain emotions (e.g love/trust) which 20th century
researchers have held to be either 'basic' or 'positive' (hope, interest, excitement) since
Darwin's list, indeed most lists, of emotions tend to include more 'negative' emotions.
9
This procedure gave us our list of 24 emotions. (We also made a set of 20 ethnographic
films which would help place these emotions in a spontaneous cultural setting and
enable us to monitor 'micro-expressions' as well as the 'rise and fall' of particular
emotions in real time). The table below, based on Ortony and Turner's pioneering survey,
presents a summary of researchers' lists of the 'basic emotions' in an initial but critical phase
of emotion research.
Theorist
Basic Emotions
Plutchik
Acceptance, anger, anticipation, disgust, joy, fear, sadness, surprise
Arnold
Anger, aversion, courage, dejection, desire, despair, fear, hate, hope, love, sadness
Ekman, Friesen, and Ellsworth
Anger, disgust, fear, joy, sadness, surprise
Frijda
Desire, happiness, interest, surprise, wonder, sorrow
Gray
Rage and terror, anxiety, joy
Izard
Anger, contempt, disgust, distress, fear, guilt, interest, joy, shame, surprise
James
Fear, grief, love, rage
McDougall
Anger, disgust, elation, fear, subjection, tender-emotion, wonder
Mowrer
Pain, pleasure
Oatley and Johnson-Laird
Anger, disgust, anxiety, happiness, sadness
Panksepp
Expectancy, fear, rage, panic
Tomkins
Anger, interest, contempt, disgust, distress, fear, joy, shame, surprise
Watson
Fear, love, rage
Weiner and Graham
Happiness, sadness
* It is noteworthy that in the enumeration above the only emotion that all the researchers have identified as
'basic' is happiness, joy or pleasure. In the few instances where happiness is not mentioned, it appears to be
substituted by 'love' or 'trust.' Our own study bears out this 'preference' for happiness as the emotion recalled
earliest by mothers and one of the most salient in terms of the narratives produced around it. Expressions of
happiness are also easily recognized across a general adult population in our study with over 90% accuracy. Our
surmise, in keeping with several other studies, is that 'happiness' supported by the release of oxytocin and allied
hormones must be critical in ontogenic development since it critically aids bonding between mother and child.
ii. MORE ON METHODOLOGY:
EMPHASIS ON INTROSPECTION AND MOTHERS' RECALL
We begin by noting an external constraint: namely, that since the duration of our grant was
three years of which the first year would have to be spent on theoretical questions, our
judgement was that we could not realistically achieve a satisfactory longitudinal video-taped
study of a large enough sample of children because such studies are labour and training
intensive and also require expensive equipment. We therefore decided on a large cross-
10
sectional study based on a simple assumption. We thought it reasonable to assume on both
biological and cultural grounds that mothers had to be evolutionarily attuned, in a
‘Darwinian’ sense, as well as skilled at being able to recognize emotions, in order to
appropriately respond to their children's wants, needs and desires. Such an ability would
significantly help ensure the survival of their children and thus would count as a powerful
genetic as well as a cultural imperative. Hence, we formulated a simple, two-part working
hypothesis.
Our hypothesis was that: a. mothers could be considered 'best observers' of their children's
growth and development in that they shared a degree of intimacy, both embodied and
cognitive, with their children that could not be matched by any researcher and would thus be
able to narratively report on emotional nuances; b. mothers had access to deep introspective
resources, especially memories of their children's development which was a reservoir we
could tap and which was, again, simply unavailable to researchers as external observers. This
two-part hypothesis would, we felt, motivate a three-stage analysis of three processes that
linked our mothers' 'recognition' data with their 'recall' and 'record' data as follows:
a. Recognition (by 505 mothers of facial expressions + 510 adult males)
b. Recall (by the same 505 mothers of their own children aged between 1-10 as to when
each of these 24 emotions appeared in their children)
c. Record (of these 505 mothers' narratives of the contexts in which some or all of these
24 emotions appeared in their children, mothers' descriptions of the facial expressions
and body language accompanying these emotions and their views on whether these
emotions were socially approved or not).
On the basis of the above schema, we first reasoned that if mothers could reliably recognize
each of the 24 emotions we showed them, this gave us strong grounds for assuming that they
would be able to identify expressions of these emotions in their own children as well as to
verbally describe the facial and bodily expressions associated with particular emotions and
introspectively narrate incidents that involved these emotions in their children. Thus, our
study was based on a detailed analysis of three types of data both quantitative and
qualitative. In devising our methodology, we bore in mind that a reiterated argument that
Ortony and Turner examine is that the basic emotions are judged to be more basic because no
11
matter what culture people come from they can recognize and identify these emotions from
facial expressions (Ekman. 1972,1990).
Our study involved the recognition of 24 emotions (with 4 photographs of each emotion
sorted into sets of: i. the prototype; ii. close to the prototype; iii. far from the prototype;
iv. random, i.e. 24x4 = 96) in a non-western but urban population. Following a pilot study,
we presented these photographs to 1015 adults (505 mothers of young children since our
main interest was in the developmental order of emotional acquisition; and 510 males = 1015
respondents).
Our results, based on one of the largest ever studies conducted on emotion recognition and
recall as well as 505 hundred hours of audio-taped narrative, are statistically robust and use
several measures over a large sample. We describe this sample briefly below:
SAMPLE: This sample of mothers was 'purposive' in the general sense that it targeted
mothers. It was quite balanced in the following ways:
1. All our mothers lived in the National Capital Region (NCR, territories around Delhi) and,
since this region attracts people from across the country, represented the diversity of India in
terms of religious and caste affiliations. The general census of the population of India has
approximately 81% Hindus, 16% Muslims, 1.00 % Jains and Buddhists, 1.5 Sikhs% and
1.00% Christians. These percentage are almost exactly reflected in our sample, except for a
slight spike amongst the Jain mothers, owing to the fact that one of our field associates is
Jain. However, these numbers can be added to the general 'Hindu' population making the
match even more exact - but even with the Jains as a separate category, this sample is pretty
representative of India in terms of religion.
500
450
400
350
300
250
200
150
100
50
0
CASTE AND RELIGION
General
Other
Total
Hindu
Muslim
Caste
Jain
Christian
Sikh
Religion
Frequency
417
55
472
365
96
33
3
8
Percent
82.6
10.9
93.5
72.3
19.0
6.5
.6
1.6
12
TABLE VII: DEMOGRAPHIC OF MOTHERS' SAMPLE
2. The sample had mothers of children in three age bands (167 mothers of children aged
below 2; 173 mothers of children below age 5;165 mothers of children between 5 and 10), so
that we would be clear about 'recency effects' in mothers' recall.
3. The number of male (273) and female (232) children is more or less evenly distributed
across the sample.
4. The number of mothers living in joint families (276), which is still a common social setup
in India even in urban areas, is about the same (slightly larger) than those living in nuclear
families (222). Seven (7) mothers did not respond to the question about the type of family to
which they belonged.
5. Almost all the mothers are Hindi speaking.
6. The mothers fall into 4 socioeconomic bands, roughly following a bell-curve.
7. The education level of the mothers in the sample follow a similar bell pattern with most
mothers being graduates but a miniscule number (21) at two ends of the spectrum who were
either illiterate (11) or had Ph.D.s (10) or similar 'high' qualifications.
8. Only a third of the mothers were working which, again, roughly reflects the situation in
urban India.
9. The average age of the mothers was 30 years and the average number of children in each
family was less than 2; this again is representative of urban Indian mothers as a whole (9
mothers were mothers of twins).
10. There were no mothers with severely disabled children in this sample because our first
priority here to build a 'typical' chart of child development in an embedded cultural context
that could then perhaps be useful for assessment and other purposes.
Most of all, we wish to emphasize that the mothers in our study, whatever their background,
were admirably articulate, and energetic in their observation of their children's behaviour. In
terms of methodology, scope and cultural setting we thus feel that our study is unique. Based
on the hypotheses and methodology outlined above, we now present our main findings.
MAIN FINDING A.
ALL ADULTS SEEM EXCELLENT AT RECOGNIZING A LARGE RANGE OF
EMOTIONS: EVIDENCE AGAINST THE 'BASIC EMOTIONS' HYPOTHESIS?
This large scale study with 1015 adult subjects seem unequivocally to show that all adults
seem to be excellent at recognizing a large emotional range with 80-90% success rates of
recognition in most cases. The 'basic emotions' are very accurately identified by both adult
males and females, but then so are the putative 'non-basic' emotions. It is true that in our
data there are some statistically significant differences between the sexes:
13
Quite a few emotions (9) are better identified amongst females while males have a higher
success rate at identifying just 2 emotions (love and contempt). However, this does not
overall affect the high success rate at identifying a large range of emotions (such a large
number of emotions - 24 - have never been tested for in any previous study; see Tables I, II &
III). In the 200 cases where we measured the time taken to identify the 'correct' emotion,
corresponding results emerged. Timing significantly correlated with accuracy (longer times
taken, more accurate identification) but the basic emotions were not identified any faster.
Recognition Accuracy 1015 Adults (510 male & 505 fem.)
105
100
95
90
85
80
75
70
65
60
55
50
45
40
35
30
25
20
15
10
5
0
Helplessness
Contempt
Interest
Hope
Obstinacy
Guilt
Courage
Excitement
Love
Boredom
Hate
Sadness
Jealousy
Shame
Happiness
Shyness
Pride
Peace
Reflection
Fear
Anger
Surprise
Laughter
Disgust
Recognition Accuracy
TABLE 1. ACCURACY OF RECOGNITION (24 EMOTIONS)
Male and female recognition
120
100
80
60
40
Male Accuracy %
20
Female Accuracy %
Happy
Sad
Fear
Anger
Surprise
Disgust
Contempt
Shame
Guilt
Obstinate
Love
Hate
Helpless
Hope
Laughter
Courage
Pride
Peace
Interest
Bore
Jealous
Excite
Shy
Reflect
0
14
Happy Surprise ContemptObstinacy Love
Hate Courage
EMOTIONS
Pride
Interest
Bore Jealous
TABLES II & III. SIGNIFICANT DIFFERENCES IN ACCURACY OF
RECOGNTION AMONGST MALES AND FEMALES
Findings and implication: A strong implication of the above findings is that the basic
emotions do not seem to emerge as 'more basic' on the parameter of recognition amongst
adults. Rather, adults seem to have the cognitive skills to 'instantly' recognize a large
repertoire of emotions, just as they can immediately process incoming sensory or linguistic
information 'online'. In short, they can 'read' a 'script' derived from permutations and
concatenations of facial muscles with ease, distinguish them from other combinations as well
as cognitively identify, and in most cases, name the relevant emotion in 'no time.' On 13 of
these emotions, men and women do equally well while men are better at reading 'love' and
contempt. Females are better than males at recognising 9 emotions, namely, happiness,
surprise, hate, courage, pride, interest and boredom - but not all that much better. Overall
both men and women are extremely good at instantly recognizing a range of emotions,
although we suspect that there might be cultural differences in reading some expressions, for
example 'contempt'.
We are keen to learn whether the test materials that we have created will show similar
results across cultures and would be happy to share our database of photographs, our
questionnaires and interview protocols with bona-fide researchers everywhere. Here are some
of the photographs we used as part of our original sample (please see Appendix for all 96
picture-sets of 24 emotions).
male
female
male
female
male
female
male
female
male
female
male
female
male
female
male
female
male
female
male
female
540.00
520.00
500.00
480.00
460.00
440.00
420.00
Male
female
MEAN RANK
Significant differences on recognition
15
HAPPINESS
SADNESS
MAIN FINDING B.
A CLEAR DEVELOPMENTAL ORDER WAS FOUND FOR THE ACQUISITION OF
EMOTION: EMPIRICAL EVIDENCE FOR THE BASIC EMOTIONS (or AN EAD)?
2. A second argument Ortony and Turner examine is that the basic emotions are basic
because other, more 'complex' emotions build on these more 'simple' building blocks. They
contend that this is an empirical question to be examined on the basis of inductive evidence.
Our study is in fact empirical insofar as it involved a large cross-sectional sample of
amenable-to-quantitative analysis, questionnaires plus qualitative hour-long audio-taped
interviews with 505 mothers of children in three age-groups (167 mothers of children aged
below 2; 173 mothers of children below age 5; 165 mothers of children between 5 and
10).
We suggest that there are at least two straight-forward empirical ways in which evidence can
be found for or against the hypothesis that there is an order from simple to complex in the
acquisition of the emotions: First, the reported or observed age of emergence of various
emotions; second, whether, from qualitative analysis, the allegedly more basic emotion
prompt more narratives from mothers and are also found embedded in narratives of the
more complex emotions.
16
For example, if the basic emotion of 'fear' is referred to more frequently in narratives of
a complex emotion like 'guilt' whereas the reverse case (guilt occurring in fear
narratives) is less frequently found, this would constitute indirect evidence for the 'basic
emotion' (EAD) hypothesis.
After we had obtained our results on our 'recognition' study which established that our
mothers, like the general population, we in fact aces at recognizing a large range of emotions,
we thus asked the 505 mothers' in our study to now recall the ages at which they saw each of
the 24 emotions they had already identified in their own children. Our method involved
requesting them to assign the ages at which they recalled seeing these specific emotions in
their children to fifteen categorical age bands. 0-3 months; 4-6; 7-9; 10-12; 13-18; 19-24;
25-30; 31-36; 37-42; 43-48; 49-54; 55-60; 61-72; 73-84; and above 85 months) at which they
first observed their own children exhibiting each of these twenty-four emotions.
The unsorted graph below shows the ages at which mothers recalled the development of
emotions in their children.
Excitement
Peace
E
M
O
T
I
O
N
S
0
5
10
Interest
Obstinacy
Surprise
Anger
Fear
15
Boredom
Pride
Hope
Helplessness
Hate
Love
Sadness
Happiness
Jealousy
Courage
Laughter
Reflection
Shyness
20
Guilt
Shame
Contempt
Disgust
25
30
35
40
45
Age at which Emotion was shown as Recalled by Mothers in Months
TABLE IV: ORDER OF EMOTIONAL ACQUISITION AS REPORTED BY MOTHERS (UNSORTED)
However, if we now sort the graph below into age bands, here is what we get:
17
Median Age Recall (ascending order)
Reflection (N=314)
Hate (N=152)
Boredom (N=343)
Pride (N=321)
Guilt (N= 353)
Shame (316)
Contempt (N=117)
Shyness (N=384)
Jealousy (N=250)
Hope (N=378)
Courage (N=389)
Helplessness
(N=321)
Interest (N=454)
Excitement (N=440)
Obstinacy (N=462)
Disgust (N=378)
23
21
19
17
15
13
11
9
Love (N=489)
Surprise (N=401)
Anger (N= 465)
Fear (N= 440)
Peace (N = 331)
Laughter (N= 477)
Sadness (N=434)
Happiness (N=500)
7
5
3
1
0
5
10
15
20
25
30
35
40
45
Age in
TABLE V ORDER OF EMOTIONAL ACQUISITION AS REPORTED BY MOTHERS (SORTED)
Once sorted, the chart above is revealing. It shows that, across 505 unrelated mothers
recalling 24 different emotions, separately interviewed by three different researchers across a
range of communities, there is a clear order in which most mothers recall the appearance
of particular emotions in their children. Further, the great majority of mothers report that
all the twenty-four emotions appear in their children by 40 months; in other words, a
negligible number of mothers use the higher age-bands after 3-and-a-half to 4 years specified
above even when they are mothers of children aged between 5-8 (approximately one-third of
the mothers). Moreover, .these emotions are recalled, after the foundation of 'happiness' is
laid, in clustered stages. Very imprtantly, it seems to emerge robustly from this study (for
the Ns are high) that the intuitively 'positive' and 'negative' emotions are actually
balanced all the way until our postulated Stages VI and VII (see Table below) when the 'selfconscious', negative emotions and 'moral emotions' all come bunched together. This can
be modelled as a stage-by-stage developmental process as below.
18
Stage I: (at or before 5 months)
Happiness (1 positive emotion paramount)
Stage II: (at 8 months)
Sadness and Laughter
(2 further emotions, 1 positive and 1 negative)
Stage III: (around 15 months)
Love and Peace, Anger and Fear
Stage IV: (around 22 months or
Surprise is read both positively and negative by mothers
(2 positive + 2 negative, 1 ambivalent = 5 emotions
added to repertoire)
Interest and Excitement, Obstinacy and Disgust
2 years)
Stage V: (around 27 months or
2 and a half)
Stage VI: (around 33 months or
almost 3 years)
Stage VII: (around 39 months or
almost 3 and a half years)
(2 positive + 2 negative = 4 emotions added)
Hope and Courage, Jealousy and Helplessness, with Shyness
once again holding the balance
(2 positive + 2 negative + 1 ambiguous emotion = 5 emotions
added)
Boredom, Pride, Guilt, Shame, Hate, Contempt
(all generally accounted negative emotions but which also
involve self-consciousness and 'moral' judgments)
Reflection
(it should be noted that, owing to the translation of this word see Appendix V - as 'thinking' or 'sochna' in Hindi which is
closely associated with the common phrase 'soch may parna' or
'to fall into thought', a fair number of the mothers in our study,
perhaps a third, saw this emotion as 'brooding' or involving
'worry' and thus negative.
TABLE VI STAGES OF EMOTIONAL GROWTH
Discussion and Implications: In general, we feel emboldened to postulate on the basis of
this large-scale body of maternal recall data something like seven stages of emotional
development. (There is, of course, nothing sacrosanct about the number seven here and we
must note that we simply followed previous research on developmental order in this respect,
see Lenneberg, 1964 and Berks, 2004 below, for example). In our view, of much more
special interest here is our finding, as mentioned, that a. evidence for a clear order of
emotional growth emerges robustly from our study; b. very few mothers in the use any of
the six higher age-bands after 3-and-a-half to 4 years even when offered the choice and even
when they are mothers of children aged between 5-8 (approximately one-third of the
mothers); c. the positive and negative emotions are evenly balanced in the early stages of
emotional development (Stages I-1II) seeming to form an (EAD) base for the more complex
culturally, linguistically imbued emotions to follow in Stages IV-VIII.
19
We should also note that it is not that such an order has not been suggested before. Laura
Berks (2004) references a number of such studies but they are quite small-scale (typically
following 10-20 mother child dyads longitudinally) and patchy, with details not available for
the research methods followed. Other more recent papers (Barrett, Lindquist and Gendron,
2007; Tracy, Robins and Tangney, 2007; Widen, and Russell, 2008; Shweder, Haidt, Horton,
and Joseph, 2008; Wierzbicka, A. 2009; Nair, 2011; Oatley 1012) have argued, in different
ways, for the close relationship between language and culture and Widen and Russell, in
particular, have contended, like us, that emotional development is gradual, like language.
However, these perspectives and studies are far fewer than expected and none have involved
the introspective memory banks of mothers as part of a large and robust cross-sectional
study or been based on a quantitative study of 24 emotions as well as narrative analysis.
We also feel, however, that wherever an introspective method relying on recourse to
memory is involved the moot question of 'subjective bias' cannot be lightly dismissed. With
regard to our data, the most obvious criticism would be that mothers, especially, are not indeed cannot be - 'objective' in their assessments of their children. They would be
influenced by both cultural and emotional factors in their responses. We might respond to this
objection by contending that all researchers are guided by cultural and emotional factors
when they assess ‘social reality’. This is so especially where children's emotions are involved
since they cannot be directly questioned and ‘self-reports’ (a standard method used in
qualitative psychological research) are impossible. Hence 'objective' researchers are as
much interpreters here as mothers - and perhaps with far less thorough, everyday
knowledge of their subjects than mothers.
Nevertheless, we are aware that the question of bias is a pivotal one that cannot be wished
away and we thus discuss two possible cases of such bias and the possible insights that accrue
from them below.
MAIN FINDING C:
A CLOSE EXAMINATION OF POSSIBLE BIASES IN MOTHERS' RECALL: FURTHER
EVIDENCE FOR THE BASIC EMOTIONS?
1. The 'Recency Effect': First, anticipating that there was likely to be a 'recency' effect in our
sample of 505 in that mothers of younger children would be likely to report seeing an
emotion earlier in their children, we had in our initial selection deliberately looked for
20
mothers whose children fell into three age bands: i. we included 167 mothers of very young
children (below 2); ii. 173 mother of slightly older children (upto 5); and 165 mothers of
older children (under 10). We reasoned that if there was such a ‘recency effec’t, we could
take a mean over our large sample of 505 mothers and still get a reliable picture of the ages at
which mothers recalled seeing a particular emotion in their children.
This in fact is what we did - but first, a graph.
12.0000
10.0000
8.0000
6.0000
Group1 (1 to 30 mnths
4.0000
Group2 (31 to60 mnths)
2.0000
Group3 (61 to 120 mnths)
.0000
TABLE VIII: THE RECENCY EFFECT IN MOTHERS' RECALL OF THEIR
CHILDRENS' EMOTIONAL ACQUISITION
For 12 of the 24 emotions, we found no significant difference in the ages at which mothers
recalled seeing the emotions in their children across the three groups. These were the
emotions of happiness, love, fear, surprise, guilt, helplessness, laughter, interest, pride,
jealousy, shyness and contempt. But, as the graph above shows, we did find significant
differences amongst the mothers in the three groups for the ages at which they recalled the 12
other emotions. Here, as expected, mothers of the youngest children consistently recalled
seeing an emotion earlier; mothers of children in the second age range recalled seeing the
emotions somewhat later; and the mothers with the oldest children reported seeing the same
emotions at a later age. So there certainly was a 'recency effect’. However, what springs
out at us from the above graph is that, even where there are significant differences in age
recall, the pattern, astonishingly, is almost exactly the same for all three age-groups.
That is, all the mothers agreed on the stage-by-stage order of the acquisition of these 24
emotions in their children, as shown in Table II even if their estimates of the ages at the
21
which these emotions appeared varied (but consistently). We consider this a strong result in
favour of the 'basic emotions' (EAD) hypothesis.
2. Non-Reporting: We also noticed another sort of 'bias' in our study: There were many
amongst our 505 mothers who said they 'did not recall' a particular emotion in their
children, especially and understandably if it was a negative emotion like 'hate' or 'contempt'.
We were especially intrigued to find the N in the above 'not recalled' sample varied
considerably from 505 mothers who all recalled their children showing happiness by the age
of 5 months to only 117 mothers who responded to the question about their children
expressing the emotion of contempt. We therefore decided to investigate this 'No Response'
or 'Not-reported' category further and here we came upon a distinct dichotomy in mothers'
recollections. There were, we found, emphatically two kinds of 'No Response'. There were
mothers who said they did not recall the age at which a certain emotion occurred and there
were mothers who said they did not notice the emotion at all in their child. So how did this
difference look in terms of numbers?
1 Not expressed (Percent)
G3 (130)
G2 (76)
G3 (125)
G2 (98)
G1 (142)
G3 (131)
G2 (116)
G1 (136)
G3 (142)
G2 (167)
G1 (109)
G3 (158)
G2 (61)
G1 (99)
G3 (124)
G2 (143)
G1 (145)
G3 (131)
G2 (145)
G1 (20)
G3 (146)
G2 (157)
G1 (145)
G1 (167)
100
80
60
40
20
0
HapySad FearAnger
Surprise
Disgust
Contempt
Shame
Guilt
Obstin
LoveHate
Helples
Hope
Laughter
Courage
Pride
Peace
Interest
Bore
Jealous
ExciteShy
Reflect
2 Don't remember(Percent)
G3 (130)
G2 (76)
G3 (125)
G2 (98)
G1 (142)
G3 (131)
G2 (116)
G1 (136)
G3 (142)
G2 (167)
G1 (109)
G3 (158)
G2 (61)
G1 (99)
G3 (124)
G2 (143)
G1 (145)
G3 (131)
G2 (145)
G1 (20)
G3 (146)
G2 (157)
G1 (145)
G1 (167)
18
16
14
12
10
8
6
4
2
0
HapySad FearAnger
Surprise
Disgust
Contempt
Shame
Guilt
Obstin
LoveHate
Helples
Hope
Laughter
Courage
Pride
Peace
Interest
Bore
Jealous
ExciteShy
Reflect
22
TABLES IX AND X MOTHERS' RECALL:
'NOT EXPRESSED' VERSUS 'DON'T REMEMBER' ACROSS THREE AGE GROUPS
Percent
Consolidated percentage of 'not expressed' versus
'don't remember'
80
70
60
50
40
30
20
10
0
1 ( Not express)
2 (doesnt remember)
Emotions
TABLE XI MOTHERS' RECALL PERCENTAGES OF 'NOT EXPRESSED' VERSUS 'DON'T REMEMBER'
It is apparent from the graph above that the number of mothers who said their child did not
show a particular emotion far exceeded the mothers who said they did not remember the age
at which the emotion was expressed. For instance, in the striking case of the emotion of
contempt approximately 70 % mothers belong to the 'my child did not express this emotion'
category and only 4% belonging to the 'I don't remember when my child first expressed this
emotion' category. A similar 'bias' is reflected in mothers’ reports of almost all the other 'not
reported' emotions.
In short, mothers overall show a strong preference not to see negative emotions in their
children; for example, in the case of disgust amongst the basic emotions, which the ‘picture
recall’ data in our study tell us is in fact the most accurately identified emotion, 22%
mothers still say they do not notice this emotion in their children. Further, this 'no response’
data also reflected interesting differences in recall between:
a. the basic emotions and the more complex emotions;
b. the positive and the negative emotions;
We present pie-charts for mothers' 'No Response' to various emotions below:
23
Colour Codes: Green: Expressed; Blue: Not Expressed; Brown: Don't Recall Age
EMOTIONS EXPRESSED AND NOT EXPRESSED
Now, here we find that something noteworthy. In the case of the basic emotions, it is
evident that even when these emotions are negative (sadness, fear, anger, disgust) mothers
systematically do report seeing them significantly earlier than other 'complex' negative
emotions (shame, guilt, contempt, helplessness and obstinacy).
BASIC EMOTIONS
HAPPINESS
HAPPINESS (5 MONTHS);
SADNESS, LAUGHTER (8
MONTHS);
1%0%
99%
FEAR, ANGER, SURPRISE,
LOVE AND PEACE (15
MONTHS);
DISGUST (21 MONTHS)
LOVE
ANGER
2%
1%
3%
5%
97%
92%
SADNESS
SURPRISE
7%7%
9%
80%
86%
FEAR
9%
87%
DISGUST
22%
4%
75%
3%
TABLE XII 'BASIC EMOTIONS' PIE CHARTS
11%
24
POSITIVE EMOTIONS
EXCITEMENT
INTEREST
LAUGHTER
1% 4%
7%6%
2%
8%
87%
90%
95%
HOPE
COURAGE
18%
77%
PEACE
28%
20%
5%
75%
5%
66%
6%
TABLE XIII 'POSITIVE EMOTIONS' PIE CHARTS
In the case of the 'positive' emotions, we also notice a clear trend. There is definitely less
'recall reluctance' by mothers - but now consider the so-called 'negative' emotions, which
actually outnumber the 'positive' by a factor of nearly 2:1.
NEGATIVE EMOTIONS
25
OBSTINACY
BOREDOM
SHYNESS
6%2%
24%
19%
5%
92%
76%
GUILT
68%
SHAME
PRIDE
26%
32%
31%
70%
4%
8%
62%
64%
6%
5%
HELPLESSNESS
33%
NOT EXPRESS
DOESN'T REM
RESPONSES
GIVEN
JEALOUSY
REFLECTION
38%
49% 47%
57%
63%
4%
5%
4%
HATE
CONTEMPT
NOT EXPRESS
24%
30%
DOESN'T REMEMBER
RESPONSES GIVEN
5%
64%
71%
6%
a.
TABLE XIV NEGATIVE EMOTIONS PIE CHARTS
The charts above show that the picture with regard to possible 'bias' in the 505 mothers who
participated in our study cannot be simply captured in terms of a 'recall reluctance' to report
negative emotions in their children. This is so because we find that, while there is indeed a
26
clear preference amongst mothers to report on displays of 'positive' emotions rather than
'negative' ones in their children, this preference is not unequivocal since:
a. the 'basic emotions' are still reported as occurring earlier than other more complex
emotions, even when they are negative
b. some 'positive' emotions such as 'hope' and 'peace' are not reported by mothers as at all
seen in their children 20% and 28% of the time respectively, which is roughly comparable to
the percentages for 'not seen in my child' negative emotions such as 'shyness', 19%; boredom
24%; and guilt, 26%. Our conjecture at the moment is that socio-cognitive factors may play
a role in the case of these more complex emotions. For example, ‘shyness’ in our sample is
judged a negative emotion to be 'socially discouraged' by educated, working mothers in the
higher socioeconomic groups but is judged a 'positive' emotion to be encouraged by relatively
less educated mothers who do not go out to work. The pattern is very similar for 'guilt'.
c. we have also discussed the case of 'obstinacy' earlier, which is a culturally powerful
construct in northern India in the form of 'zidd' and suggest that 'boredom' may be part of this
semantic continuum as an 'emotion' associated with children and therefore felt by the mothers
in our study to be expected and tolerated in their own children.
d. the complex often called ‘self-conscious’ emotions, especially if negative, tend not to be
reported; moreover, there is a sort of 'moral' gradation here with guilt (not reported by 26%
of mothers in their children); pride (not reported by 31%); shame (not reported by 32%);
helplessness (not reported by 33%); jealousy (not reported by 47%); reflection (often
interpreted as 'worry' or 'brooding' by mothers and seen last but still more positive at 57%
than the emotions which follow), namely, hate (not reported by 64%); and contempt (not
reported by 71% of mothers). These results are tabulated below.
percentage of 'moral' emotions found by our 505
mothers not to be expressed by their children
100%
0%
TABLE XV: 'MORAL' EMOTIONS REPORTED BY MOTHERS AS 'NOT EXPRESSED' BY THEIR CHILDREN
27
Discussion and Implications: Our suggestion is that the high percentages of 'non-reports' of
certain self-conscious emotions shown above indicate an implicit 'Moral Scale' that mothers
operate with. We also conjecture that this implicational scale of: Guilt->Pride->Shame->
Helplessness -> Jealousy -> Reflection-> Hate -> Contempt could well be judged by the
mothers in our study as part of an 'adult' emotional repertoire that robs their children of their
childhood and so they are less than keen to attribute these emotions to their children. It is
noticeable that quite a few mothers in fact feel that some amount of guilt, pride, shame and
helplessness is acceptable and even perhaps to be encouraged in their children but no
mothers admitted that hate and contempt were acceptable emotions. Reflection, as
mentioned, was a complex and ambiguous semantic concept that our mothers interpreted in
very varied ways. They were, however, good at recognizing it in the picture recognition task.
The tables below show the percentages of 'positive' and 'negative' emotions reported by
mothers as 'not expressed' in their children. The 'basic emotions' are shown in red.
Percentage of positive emotions not expressed
30%
20%
10%
0%
TABLE XVI: PERCENTAGE OF POSITIVE EMOTIONS NOT EXPRESSED
Percentage of negative emotions not expressed
80%
60%
40%
20%
0%
TABLE XVII: PERCENTAGE OF NEGATIVE EMOTIONS 'NOT EXPRESSED'
28
Once again, the percentage charts of emotions reported by mothers as 'not expressed' by their
children are revealing for the following reasons:
a. a clear 'preference trend' is shown for positive emotions being seen by mothers in their
children: for example, if we consider the top 'not reported’ positive emotions, they stand at
approximately 16% for courage, 18% for hope and 27% for peace, whereas the negative
emotions are 'not reported' at almost 52% for reflection (mostly interpreted as 'worrying' or
'brooding' by mothers), 61% for hate and 70% for contempt: these differences are stark.
b. in the case of both positive and negative emotions, the 'basic emotions' are all found to
appear significantly at the lower left hand corner of the 'not-reported' percentage scale
(0%-10% percent for the 'basic' positive emotions and 1%-20% for the 'basic' negative
emotions)' whereas the non-basic emotions appear on the right-hand higher end of the scale
(6% to 27% approximately for the non-basic positive emotion and 18% to 70% for the nonbasic negative emotions: this constitutes, in our view, strong evidence for the basic emotions.
c. these percentages for non-reported emotions correlate well with both order of emotion
acquisition (see Table V above) and with the 'narrative space' accorded by mothers to the
basic versus non-basic emotions (see Tables XIX to XXV below).
We will now attempt to relate these findings to a foundational template for developmental
patterns in children suggested by Eric Lennenberg.
MAIN RESULT D:
LENNENBERG REDUX: THE CO-DEVELOPMENT OF LANGUAGE & EMOTION
As early as 1964, Lennenberg created a chart for stages of language development which has
been augmented by several detailed studies of language acquisition since (see Lust, 2008).
However, emotional development has been relatively little studied and most studies have
involved smaller groups such as 10 to 20 mother-child dyads (see Berks, 2004). In our
research we have created a parallel column to the 'motor development' and 'vocalisation and
language development' stages suggested by Lennenberg. This is a partial version of that chart
in the sense that, although, we have detailed analytic descriptions of each one of the 24
emotions that we have studied, it is unwieldy to include them all here, so at the moment. we
have included in red all the stages of development and detailed observations on only two
contrasting emotions (happiness and contempt), just to give a flavour of our enterprise.
29
Motor Development


12 weeks
Supports head when in
prone position; weight is on
elbows; hands mostly open;
no grasp reflex
Vocalization and Language




 16 weeks
Plays with a rattle placed in his
hands (by shaking it and staring at
it), head self-supported; tonic neck
reflex subsiding


20 weeks
Sits with props




6 months
Sitting: bends forward and
using hands for support;
can bear weight when put
into standing position, but
cannot yet stand with
holding on; reaching:
unilateral; grasp: no thumb
apposition yet; releases
cube when given another




8 months
Stands holding on; grasps
with thumb apposition;
picks up pellet with thumb
and finger tips




10 months
Creeps efficiently; takes
side- steps, holding on;
pulls to standing position


12 weeks
Markedly less crying than
at 8 weeks; when talked to
and nodded at, smiles,
followed by squealinggurgling sounds usually
called cooing, which is
vowel like in character and
pitch-modulated; sustains
cooing for 15-20 seconds
16 weeks
Responds to human sounds
more definitely; turns
head; eyes seem to search
for speaker; occasionally
some chuckling sounds
20 weeks
The vowel-like cooing
sounds begin to be
interspersed with more
consonantal sounds; labial
fricatives, spirants, and
nasals are common;
acoustically, all
vocalizations are very
different from the sounds
of the mature language of
the environment
Emotional Development
(our findings)
** 12 weeks: Approximately 200
mothers in our study, recall seeing
'happiness' in the age-band 0-3
months (i.e. at or before 12 weeks).
However, since we calculated the
mean recall age by 505 mothers in
total, this brought the mean age at
which the mothers' in our study
recalled seeing happiness to about 5
months (see below).


20 weeks
(Happiness)

Some frequent terms
mothers use to describe
'happiness': smile, hug,
jump, happy face, clapping

Excerpt from a mother's
description of happiness in
her child, Pratyaksh:
Interviewer : toh kya baat thi - kuch
aisa yaad hai ki kis baat pe khush hua
tha? So do you remember why he was
happy?
Mother : Starting mein aiski khaas
baat jo mujhe yaad hai ki apne aap
baitha hua hasta rheta tha, matlab bahut
hasmukh tha. Dekh ke sab yhe khete
the ki aisa baccha bhagwan sabko de.
Mother- in-law us time thee nahin aur
ye chota tha , toh mein razaee charon
taraf laga ke sardiyon mein bhitha deti
thee, toh muskurana kisi baat pe , kuch
8 months
kar rahein hain toh bhee muskurana aur
Reduplication (or more
kabhee kisi baat pe pareshaan nahin
continuous repetitions)
karna matlab bahuuut hee shant
becomes frequent;
intonation patterns become baccha kehna chahiye.
distinct; utterances can
Translation: In the beginning, the
signal emphasis and
main thing I remember is that he
emotions
kept laughing - he had a very jolly
nature! Whoever saw him used to
say that everyone should have a baby
10 months
like him . He was very young and my
Vocalizations are mixed
mother -in law was not available
with sound-play such as
then, so I just made him sit under
gurgling or bubblethe sun in the winter and wrapped a
blowing; appears to wish to blanket around him but he kept on
imitate sounds, but the
smiling whatever we did and never
imitations are never quite
troubled me at all. I must say he was
successful; beginning to
a very peaceful kid.
differentiate between words Interviewer : agar expressions ke bare
mein baat karoon, toh muskurahat
heard by making
dikhti thee face pe, aur kya kya
differential adjustment
6 months
Cooing changing into
babbling resembling one
syllable utterances; neither
vowels nor consonants have
very fixed recurrences;
most common utterances
sound somewhat like ma,
mu, da, or di
30
expression the , jab khush hota hai toh
aur kaise expressions hote? If I talk
about the expression 'smile', it shows
on his face. What other expressions
does he show when he's happy?
Mother: abhee ke bhi bata sakte hain
na? Can I talk about his present
expressions as well?
Interviewer: haan haan , abhee ke bhi
bata sakte hain aur phele ke bhi ki kya
kya kiya. Yes. Yes, you can talk
present and what he did earlier too.
Mother : haav bhav mein toh chehere
ka jaise batayen toh khul ke hasta hain
If I am to talk of his behaviour, his
countenance, he laughs in a most
open fashion.
Interviewer : muskurahat hoti hai
chahere par, aur koi antar dhikta hai?
Apart from his smiling demeanour,
do you see other things?
Mother: haan, kabhee kabhee mere se
aake chipak jata hain, jab jyada khush
hota. Yes, he sometimes comes and
clings to me (hugs me) when he's
especially happy.
Interviewer: Does he say something?
Mother: No, no really.
12 months
Walks when held by one
hand; walks on feet and
hands – knees in air;
mouthing of objects almost
stopped; seats self on floor




18 months
Grasp, prehension and
release fully developed; gait
stiff, propulsive, and
precipitated; sits on child’s
chair with only fair aim;
creeps downstairs
backward; has difficulty
building tower of 3 cubes
12 months
Identical sound sequences
are replicate with higher
relative frequency of
occurrence and words
(mamma or dadda) are
emerging; definite signs of
understanding some words
and simple commands
(show me your eyes)


24 months
Runs, but falls in sudden
turns; can quickly alternate
between sitting and stance;
walks stairs up or down,
one foot forward only


18 months
Has a definite repertoire of
words – more than three,
but less than fifty; still
much babbling but now of
several syllables with
intricate intonation
pattern; no attempt at
communicating
information and no
frustration for not being
understood; words may
include items such as thank
you or come here, but there
is little ability to join any of
the lexical items into
spontaneous two-item
phrases; understanding is
progressing rapidly.
24 months
Vocabulary of more than
50 items (some children

8 months (Laughter and
Sadness)

15 months
(Love, Surprise, Anger,
Fear, Peace)

21 months
(Disgust, Obstinacy,
Excitement, Interest)

27 months/2years 3months
(Helplessness, Courage,
Hope, Jealousy and
Shyness)

33 months/2years 9months
(Contempt, Shame, Guilt,
Pride , Hate and Boredom)
Excerpt of sample narrative by
mother for the emotion of contempt:
Mother : Pratyaksh mein meine kabhi aisa
kuch nahin dekha, kafi alag sa
baccha raha hai toh, jaise jaise bada hota
gaya thoda bahut laga gussa hai
magar chotepan mein kabhi nahin tha iske
andar. Accha lekin nicha
dhikhana ya kisi ki ninda karna, aisa nahin
hai, halka phulka kabhi kuch
kah deta hai . Ab khel mein iski kamee
31




30 months
Jumps up into air with
both feet; stands on one
foot for about two seconds;
takes few steps on tip-toe;
jumps from chair; good
hand and finger
coordination; can move
digits independently;
manipulation of objects
much improved; builds
tower of six cubes
3 years Tiptoes three
yards; runs smoothly with
acceleration and
deceleration; negotiates
sharp and fast curves
without difficulty; walks
stairs by alternating feet;
jumps 12 inches; can
operate tricycle
4 years Jumps over rope;
hops on right foot; catches
ball in arms; walks line
seem to be able to name
everything in
environment); begins
spontaneously to join
vocabulary items into twoword phrases; all phrases
appear to be own creations;
definite increase in
communicative behavior
and interest in language


30 months
Fastest, increase in
vocabulary with many new
additions everyday; no
babbling at all; utterances
have communicative intent;
frustrated if not understood
by adults; utterances
consist of at least two
words, many have three or
even five words; sentences
and phrases have
characteristic child
grammar, that is, they are
rarely verbatim repetitions
of an adult utterance;
intelligibility is not very
good yet, though there is
great variation among
children; seems to
understand everything that
is said to him
kahoon ya kya lekin isko hamesha
jeetne ki icchha rheti hai aur wo nahin milti
toh phir yhe pareshaan
hota hai. Kabhee Kabhee doosren bacchho
ke liye yhe khe deta hai ki
dekha na tu khe raha tha first aaonga lekin
tu toh har gaya na, mein first
aa gaya. Lekin waise kabhee kisi ko neecha
nahin dhikhana kyounki hum
sahi bataye toh hum nirankari mission se
jude hue hai, waha main
cheez jab hum satsang mein jate hai toh
bacchho ko bhi leke jate hai .
Wahan yahee shikhaya jata hai ki kabhee
kisi ko neecha nahin dhikhan,
ninda nahin karna ya chugli nahin karni toh
isme yhe toh kabhee nahin
aya ki kisi ko neecha dhikhana hai. Anjane
mein kabhee bol de toh
pata nahin magar iske andar wo cheez nahin
hai.
Translation: I have never seen
anything like this in him. He's been a
very different kind of kid from the
time he was a child. I think as he
began to grow, he did show a bit of
anger but there was no anger in him
when he was a baby. But putting
someone down or insulting them is
not in his nature though he may
sometimes say something just a little
harsh when he gets into an
argument. So shall I tell you about
his weakness when it comes to
playing games? He always wants to
be a winner, so if he doesn't achieve
this desire, he becomes frustrated.
Sometimes he taunts other kids,
saying, hey you said you'd come first
but you lost and now I have come
first. But he has never put anybody
down because actually we are
attached to a 'Nirankar' (Sikh
religious sect) mission and always
take them (her children) along with
us to the gatherings (satsang). There,
they teach us how to avoid
contemptuous behaviour - never
treat anybody as beneath you, don't
carry tales, humiliate nobody. That's
why he's never shown an inclination
to treat people as beneath him. If
says something without [my]
knowing it, then I cannot say - but he
really doesn't have this feeling in
him.
3 years and above
Understands complex
language, similes,
metaphors and wordplay
3 years and above
(above 39 months)
Emotion shown:
Reflection
TABLE XVIII 'POST-LENNENBERG', 2012
32
Findings and implication in terms of mothers’ recall, recognition and reports of their
children's emotional growth: As is apparent from the graphs and charts presented above,
our study shows a strong order of the emergence of the emotions in terms of mothers'
memories, with the basic emotions emerging early and the more complex 'moral' and ‘selfconscious’ emotions emerging last. This finding implies that the former could indeed be the
basis for the more 'social' constructions of the latter. 'Happiness' emerged first, according to
all the 505 mothers in our study and 'reflection' at the very end over a period of roughly 0-4
years, which corresponds with our overarching hypothesis that, if there is a sort of 'module' or
EAD (an Emotional Acquisition Device) that deals with emotional processing, it is different
in developmental terms from the other faculties of touch, taste, smell, hearing and sight
which are fully functional by roughly 12 months of age in that it is 'slow-growing'. In this
respect it is more like the LAD (Language Acquisition Device) that has long been postulated
for language and develops, our results seem to show, over more or less the same period
of four years. We conjecture from this that, while a putative EAD interfaces with all the
other sensory systems just as the LAD does, it is marked in particular by a co-development
with language cognitively as well as culturally.
Our research findings seem to show that language and emotional development occur in close
tandem age-wise according to evidence from mothers' recall. It should of course be obvious
that we are not claiming here that the emotions have a parallel structure to that of language for, in any case, what would be the point of two systems or modules (the EAD as well as
the LAD) that did the same cognitive work? Our position is, rather, that emotions and
language provide the human species with high-level cognitive channels that reinforce as well
challenge existing cultural constructs and interactively help imagine every individual self as a
community participant. In this sense, both language and the emotions are empathy
generating cognitive machines.
So how do we interpret these striking results from a large-scale cross-sectional study which
has demonstrated strong abilities amongst a general (non-western) population to 'instantly'
recognize a large range of emotions and also shown that mothers recall a pattern of emotional
acquisition in their children that very clearly privileges the 'basic' emotions and displays, as
well, a complex understanding of the cognitive roles played by the self-conscious or 'moral'
emotions ? This question brings us to Ortony and Turner's last point concerning 3. Language
and Naming and to a central point of discussion in our own investigations.
33
MAIN DISCUSSION D: THE RELATIONSHIP BETWEEN LANGUAGE AND
EMOTION OR THE ‘LAD’ AND THE PUTATIVE ‘EAD’
Ortony and Turner interestingly argue that if emotional development is compared to language
development, this militates against the 'basic emotions' view, since no one after all argues that
that there are 5 or 6 basic languages on which other more complex languages are founded.
Rather, they say, language development is based on an order of development that involves
the same components across languages. These components are the elements of language
such as phonemes, morphemes, words, phrases and finally sentences. Facial expressions, for
example, would on this view, give us something like the 'phonemes' of emotion, which
would then combine into meaningful morphemes and then into larger structures. But
while this is an attractive view (with maybe the 53 facial muscles corresponding to a large 53
phoneme language repertoire!), there are two questions to be asked here, based on Ortony and
Turner's language analogy-cum-componential analysis perspective.
One, language famously has 'duality of patterning' - that is, different sequences of sound
combine 'arbitrarily' to give us different meanings (although with certain constraints of a
'universal' nature such as the rules for syllable formation). However, no one had ever
maintained that if a child is say born deaf, and therefore, dumb from birth, he or she does not
acquire language. Meaning and knowledge of meaning as well as syntactic understanding
can, in short, be acquired independently of whether or not one has physical access to
phoneme sequences. On this analogy, we can ask if there is 'duality of patterning' in terms of
the pairing of facial expressions/body language with particular emotional interpretations as
well? If there are impairments in these areas, one must then argue that a full emotional
repertoire will/can still be acquired. And where might we look for universal constraints on
the combinatorial properties of body language/facial expressions? There is scope in our
data, especially our films, for further and deeper research here.
Two, language also had a 'feedback loop', whereby we are able to 'correct' ourselves when
we make slips in language because we can hear ourselves speak. But since we cannot observe
our own facial expressions, how do we know when we are expressing things 'wrong', i.e. the
'display grammar' of our emotions is not 'making sense' to others?
Our hypothesis, given our strong evidence for a developmental order of the emergence of
emotions, is that the 'basic emotions' function like a core semantic vocabulary that is, as a
34
norm but not a necessity, mapped on facial/bodily expressions which provide the
'components' of these emotions. Just as a child is likely to acquire conceptually simple
and phonemically tractable words like ball, mama, doggie, moon etc. before it acquires
words like ominous, septuagenarian or howitzer, so too with the emotions.
Or to change the analogy for a moment: think of a cat, puppy or monkey. The 'faces' of
these animals do not have the kind of musculature that enable them to produce a variety of
expressions (just as their vocal chords do not allow for the range and permutations of sounds
humans can produce but still permit vocal communication). However, this does not
necessarily or logically mean that animals do not feel complex emotions. It is just that the
emotional ranges that animals exhibit could be different from (more limited than?) ours
which seem to permit a range of self-conscious emotions such as guilt, shame and
contempt. The difference between babies and animals is that babies seem to attain the range
of adult emotional states of the species by age 3 or 4. This is what our research seems to have
shown. We do not know about animals.
Our recognition studies demonstrate that adults are good at instantaneously processing both
simple and complex emotions, in the same way that an adult would not linguistically have
any more trouble processing the word 'therapeutic' the word 'toast.'
However, our recall studies show that mothers robustly and consistently recall the 'basic'
emotions occurring earlier in their children. This order of acquisition is also confirmed by our
narrative data which record particular emotions in cultural contexts and event frames. This
could be because, in evolutionary terms, the basic emotions were simpler for mothers
and others to decode in children since they just used fewer/more cognitively salient
facial muscles - or were less 'internalized' or autonomic than the more complex
emotions that needed words to be externalized. This is a set of empirical conjectures that
still need to be tested but studies so far by Ekman et al using the FACS (facial action coding
system) seem to indicated that this is indeed the case and our own recorded data on the stories
told by mothers about the simple versus the more complex emotions seem to bear this
conjecture out.
Another interesting characteristic of the emotions that we seem to have noticed in terms of
the language analogy is that a single emotion seems to be at one and the same time nounlike, verb-like, adjectival/adverbial as well as prepositional, i.e. 'directed'. So, for
example, fear is at once a state (noun-like), a process (verb-like), evaluated (adjectival, a
35
bad feeling) and emanating from some source (directional, prepositional). Thus, our idea is
that each 'basic' emotion contains within itself a flexible 'syntax'.
Over time, we suggest that this syntax, structurally 'projects' to other emotions that are more
complex and culturally indexed. Thus, surprise, a 'basic emotion' projects to emotions like
excitement, interest and boredom (read as a negation of interest, of the very possibility of
surprise etc.), while the moral or 'self-conscious' emotions (shame, guilt, contempt, pride)
require a modal component (should, ought, etc.). A very complex emotional canvas with
hundreds of moods and emotions can eventually be created but this is exactly where
culture, language, aesthetics and naming are critical. In effect, there are possibly just as
many emotional repertoires as there are cultures and sub-cultures but they all rely on a given
developmental sequence of emotion acquisition, socially acquired through intimate contact
with their mothers/care-givers whose own introspective states and internal narratives play a
significant part in moulding children's emotional development . As Gopnik (2009) puts it:
Babies learn about the world based on what they see their parents do, and they act
based on that knowledge. Those actions influence what their parents do, which
influences what the babies do and how they act, and so forth. A naturally sad baby
observes a sad mother, concludes that sadness is the human condition, acts sadly and
makes her mother even more sad.
We now move on to a quick description of the qualitative narrative data that we examined
in support of the hypotheses outlined above about the 'contexts of intimacy' in which
mothers' observations of the development of their children's emotions are situated.
MAIN DISCUSSION E: NARRATIVE DATA & THE 'EMOTIONAL SPEECH ACT':
505 mothers were interviewed on the development of emotions in their children of which 210
narrative interviews have been coded in detail). In the particular study reported here, we
focus on speech-act verbs in these narratives as one possible area in which we might, as an
initial research strategy, explore not so much the ‘interfaces’ between any putative 'modules'
of language and emotion strictly defined, but rather discourse interactions between what
Hauser et al (2002) call 'the faculty of language broadly defined' or FLB and 'the faculty of
emotion broadly defined' or what we call the FEB.
In this connection, Nair (2002, 2011) has argued that while the postulate of Speech Act
Theory (Austin, 1962; Searle, 1969) that one states something merely by declaring: ‘I state
36
that X’; one promises merely by saying ‘I promise that X’ etc. is widely accepted, this
observation does not hold true of a large class of emotion-verbs. It is difficult, if not
impossible, to get someone to be surprised, amazed, frightened, simply by uttering “I surprise
you!’, ‘I amaze you!’ or ‘I frighten you!’.
Emotional arousal, in short, must be linguistically achieved through indirect means. For
instance, to frighten someone, you must tell her a ghost story; to amaze someone you have to
present her with some situation that challenges her ordinary everyday beliefs, and so on. In
indirect support of this conjecture, it is found that in order to describe their children's
emotions, the mothers in our data-set almost always take recourse to a narrative, as if an
emotion cannot be described outside this discourse structure; conversely, the same mothers
report an array of affective sensations and actions (hearts thudding, screaming, throats going
dry, smiling, gasping) while engaging in story-talk with their children even as they
acknowledge that this story-experience is in fact 'false'. In our work, we discuss the
implications of these qualitative findings, suggesting that the 'extra' cognitive work
required to interface with the FEB is undertaken by the FLB via discourse structures
like narrative. Narrative, in particular, arguably has strong 'recursive' properties of the sort
Chomsky avers are central to language, functioning a bit like a ‘flight simulator’ in that it
enables co-constructed rehearsals of love, fear, anger etc. in appropriate contexts without
having to be in those crisis situations. Thus, the 'faculties' of language and emotion may
usefully co-evolve in children to ensure cultural survival in complex social worlds. Here
are some snapshots of our of our narrative data on the emotions.
MAIN RESULTS E: NARRATIVE GRAPHS
NARRATIVE DATA GRAPHS: MAXQDA
210 MOTHERS: NUMBER OF NARRATIVES
GENERATED BY EACH EMOTION TOTAL 765
37
TABLE XIX
frequency
65-35
NARRATIVES
anger
44
fear
35
happiness
65
sadness
42
shame
40
surprise
Total
36
262
TABLE XX
frequency
34-25
boredom
29
courage
30
disgust
33
excitement
27
guilt
33
helplessness
31
hope
26
jealousy
26
laughter
29
love
33
obstinacy
33
peace / calm
25
pride
34
shyness
Total
EMOTION
32
421
RANGE: 25-34 NARRATIVES
TABLE XXI
EMOTION
RANGE 15-24 NARRATIVES
TABLE XXII
38
NARRATIVE
CODES ASSOCIATED WITH THE BASIC EMOTIONS
250
200
150
100
50
0
anger
disgust
fear
happiness
NARRATIVE CODES: BASIC EMOTIONS
sadness
surprise
TABLE XXIII
LEAST NARRATIVE CODES
MOST NARRATIVE CODES
250
200
150
100
50
0
NARRATIVE CODES: MOST NARRATIVISED EMOTIONS VS LEAST
TABLE XXIV
NUMBER
OF CODES ASSOCIATED WITH
EACH EMOTION TOTAL 839
250
200
150
100
50
0
TABLE XXV NARRATIVE CODES FOR ALL EMOTIONS
39
BUT WHAT ARE NARRATIVE CODES?
In the Max QDA (Qualitative Data Analysis) system, one researcher trained in analytically
using the system teaches another and then they code a few narratives/pieces of audio or
written discourse together until there is satisfactory inter-coder reliability. In our case, we
had two such coders and they decided, in consultation with Nair (the PI), what the most
suitable coding analysing would be, given our particular concerns in this project.
We finally settled on a simple coding structure of 'descriptive' (specifically, descriptions of
facial expressions and 'body language' that occur in mothers' narratives of expressions of
emotions in their children) versus 'thematic' codes (themes that occurred/recurred in
mothers' narratives about various emotions). This would, we reasoned, give us a straightforward narrative structure for coding linguistic observations of the emotions. There is, of
course, much greater scope for more nuanced analysis of these narratives in future. For
example, since ours is high quality audio data, we could code for pitch, intonation contours
and so forth. However, it should be emphasized that these qualitative 'counts' are no
substitute for detailed, painstaking qualitative analysis such as is possible using a
conversational analytic (CA) approach. We are currently in the process of transcribing and
translating our data which consists of hour-long audio-recorded conversations with each
mother, so that we can embark on such a interactional analysis. A simplified version of
excerpts from such a transcript has already been shown in the 'Emotional Development'
column of the Lennenberg chart above. Meanwhile, here are a few examples of what we
mean by descriptive and thematic narrative codes.
SOME















DESCRIPTIVE CODES FOR EYES, FACE, LOOKING
closed eyes
strange eyes
blink eyes
bulging eyes
lowered eyes
smaller eyes
wide open eyes
glittering eyes
face angry
face broad
face contorts
face down
face frustrated
face set
face silent














good face
pouts face
red face
face strange expression
swollen face
look away
look down/ look down
quickly/ look downward
look happy
look innocent
look of achievement
look of helplessness
look scared
look (a)side
look up
TABLE XXVI DESCRIPTIVE CODES IN MOTHERS' NARRATIVES
40
SOME













THEMATIC CODES (FROM THE ‘F’
fairytales
family
family books
family romances
family, housekeeping
fear
feeling sorry, animals
feeling sorry
female films
fun stuff
finance fingers
first noticed
follow
food














& ‘G’
LISTS)
drop food
food on street
frequent
frustration
future
games
gender
general knowledge
generosity
good performance
good thing
graduate
guilt/guilty
gutter
TABLE XXVII THEMATIC CODES IN MOTHERS' NARRATIVES
UPSHOT: CAN THE COGNITIVE STRUCTURE OF AN 'EMOTIONAL ACT' BE
INFERRED FROM MOTHER'S DESCRIPTIONS OF THEIR CHILDREN'S EMOTIONS ?
Consider the following as a first approximation:
 Speech Act Structure:
F (P=A+P) i.e. Force (Proposition = Argument + Predicate)
 Emotion Act Structure: E (N=D+T) i.e. Emotion (Narrative = Description + Theme)
 So a typical fragment of a narrative structure from which we infer that a mother infers, let’s say,
anger, on the part of her child, could look something like:
E anger
(N=
D
Anger (N = Face is red
EMOCUTIONS?
EX (N = D + T)
WOULD, ON THIS
ACCOUNT,
TENTATIVELY
COMPRISE THE
COGNTIVE
SEMANTICS/
PRAGMATICS OF AN
EMOTIONAL ACT/
EMOCUTION AS IT
NARRATIVELY
EMERGES FROM
THE INTERACTION
OF THE PUTATIVE
‘MODULES’ FLB AND
FEB
+
+
T)
Toys (are destroyed))
THE BASIC
EMOTIONS
(HAPPINESS,
SADNESS, FEAR,
ANGER, DISGUST,
SURPRISE?)
WOULD BE THE
FOUNDING
‘CLASSES’ OF
EMOTION ACTS
OR ‘EMOCUTIONS’
ON WHICH MORE
COMPLEX
EMOTIONAL ACTS
ARE FOUNDED
(THE CAD
HYPOTHESIS,
PARROT, 2011
ETC.)
41
TABLE XXVIII A POSSIBLE STUCTURE FOR THE 'EMOCUTIONARY ACT'
MAIN FINDINGS F:
THE CULTURAL CONTEXT OF MOTHERS' RECALL & NARRATIVES OF EMOTION
In addition to our hypothesis that narratives in general are structurally and emotionally
crucial to an enriched understanding of emotional acts, we also analysed the
socioeconomic and cultural context of emotional acts. Here are some graphs on this aspect of
our study.
4. Emotions and Cultural Cues: Some Findings
RECALL:
M
E
A
N
R
275
N=192 N=250
N=207
250 Significant Differences Between Joint
N=172 N=242
& Nuclear
Families
N=176
N=167
N=223in Recall
225
N=165 N=197
N=162
N=148
N=142
N=205
N=183
200
N=175
N=177
175 N=170
150
125
100
75
50
25
0
TABLE XXIX MOTHERS' EMOTIONAL RECALL IN JOINT VS. NUCLEAR FAMILIES
*From the graph above, we see that mothers in Joint Families, as opposed to mothers in
Nuclear families, consistently recall earlier ages for when they spotted a number of
emotions in their children. This effect is pronounced in the case of some 'moral'
emotions like shame, guilt and disgust.
SIGNIFICANT DIFFERENCES BETWEEN SES GROUPS IN PICTURE RECOGNITION
Sad
Low (189)
High (180)
Low (188)
High (180)
Low (187)
High (174)
Low (186)
High (175)
Low (188)
High (177)
Low (185)
High (177)
Low (188)
High (180)
Low (188)
High (174)
Low (189)
High (178)
Low (188)
High (178)
Low (188)
High (178)
300.00
200.00
100.00
.00
Fear disgust love pride peaceinterest bore jealous shy reflect
42
TABLE XXX SIGNIFICANT DIFFERENCES: ALL SOCIO-ECONOMIC GROUPS IN THE
PICTURE RECOGNITION TASK
* This graph shows that mothers in Lower Socioeconomic Groups are significantly
better at recognizing: sadness, fear, love, boredom, jealousy, shyness and reflection.
Higher SES Groups better at recognizing disgust, pride, interest and peacefulness.
Accuracy of identification is the same for both SES groups for the 13 other emotions.
250.00
200.00
150.00
100.00
50.00
.00
Low (186)
HighLow
(151)
(161)
HighLow
(128)
(175)
HighLow
(180)
(121)
HighLow
(138)
(181)
HighLow
(141)
(126)
HighLow
(120)
(169)
HighLow
(165)
(153)
High (78)
Happy Sadsurpriseguilt
obstinacy
lovehelplesshopelaugh
couragepridepeace
interest
excite shyreflect
TABLE XXXI
* Mothers in Higher SES Groups consistently recall various emotions as occurring at an
earlier age in their children.
300.00
200.00
100.00
.00
Low High
(178)(167)Low High
(179)(172)Low High
(188)(182)Low High
(172)(169)Low High
(179)(174)Low High
(165)(166)
shame
guilt
laugh
jealous
shyness
reflect
TABLE XXXII
* Mothers in Lower Socioeconomic Groups feel that shame, guilt and shyness should be
encouraged in their children; Mothers in Higher Socioeconomic Groups feel laughter,
jealousy and reflection should be encouraged in their children. For all the 19 other
emotions the trends are the same - i.e. 'positive' emotions are more encouraged than
negative ones like hate and contempt.
43
TABLE XXXIII SIGNIFICANT DIFFERENCES IN MOTHER'S EDUCATION
Fear
Courage
Pride
Atte…
Gra…
Post…
Atte…
Gra…
Post…
Atte…
Gra…
Post…
Atte…
Gra…
Post…
N=156
N=157
N=155
N=179
N=156
N=251
N=178
N=181
N=153
N=156
N=152
N=179
N=167
N=141
N=150
Atte…
Gra…
Post…
500.00
.00
Peace Interest
TABLE XXXIV: SIGNIFICANT DIFFERENCES RELATING TO MOTHERS'
QUALICATIONS AND THE PICTURE RECOGNITION TASK
* Main finding: Fear is recognized more accurately by mothers with lower educational
qualifications
Postgra…
Gradua…
Attend…
Postgra…
Gradua…
Attend…
Postgra…
Gradua…
Attend…
Postgra…
Gradua…
Attend…
Postgra…
Gradua…
Attend…
Postgra…
Gradua…
Attend…
Postgra…
Gradua…
Attend…
300.00
200.00
100.00
.00
Attend…
RECALL:
Happy
SadFear
anger
surprise
shame
guilt
obstinacy
lovehate
helpless
hope
laugh
courage
pride
peace
interest
bore
jealous
excite
shyness
reflect
TABLE XXXV SIGNIFICANT DIFFERENCES RELATING TO MOTHERS'
EDUCATIONAL QUALICATIONS AND RECALL
*The more educated the mothers are, the earlier the ages at which they identify an
emotion in their children.
sadness
guilt
laugh
interest
shy
Attended school…
Graduate (156)
Postgraduate &…
Attended school…
Graduate (170)
Postgraduate &…
Attended school…
Graduate (178)
Postgraduate &…
Attended school…
Graduate (181)
Postgraduate &…
Attended school…
Graduate (180)
Postgraduate &…
300.00
250.00
200.00
150.00
100.00
50.00
.00
Attended school…
Graduate (169)
Postgraduate &…
44
reflect
TABLE XXXVI
SIGNIFICANT DIFFERENCES RELATING TO MOTHERS' EDUCATIONAL
QUALICATIONS AND ATTRIBUTIONS OF SOCIAL SIGNIFCANCE
* Main Finding: Guilt and shyness encouraged by less educated mothers
Frequency and Percentage Graph of Mother's Occupation
450
400
350
300
250
200
150
100
50
0
Frequency
Percent
Nonworking
382
75.6
working
123
24.4
TABLE XXXVII MOTHERS' OCCUPATION: WORKING/NON-WORKING
* Main Finding: Only about a third of the 505 mothers in our sample 'work' in the sense
of earning an income and they tend to belong to higher socioeconomic groups. This
roughly reflects the situation on the ground in urban India at present.
45
SIGNIFICANT DIFFERENCE OF WORKING & NON
WORKING MOTHERS ON PIC. RECOG. TASK
Fear
Helpless
Peace
Bore
Jealous
Working (121)
Nonworkin (379)
Working (118)
Nonworkin (377)
Working (119)
Nonworkin (373)
Working (119)
Nonworkin (377)
Working (119)
Nonworkin (376)
Working (120)
Nonworkin (378)
300
250
200
150
100
50
0
Reflect
TABLE XXXVIII SIGNIFICANT DIFFERENCE OF WORKING & NON WORKING
MOTHERS ON PICTURE RECOGNITION TASK
* Non-working mothers recognize fear, helplessness, boredom, jealousy and reflection
more accurately. This reflects the pattern found in Lower SES Groups and in mothers'
with less educational qualifications, suggesting that some mothers (low income, less
qualified and non-working) are emotionally affected by negative social circumstances.
Happy Surpris Shame
Love Helples Laughtr Pride
EMOTIONS
Nonworkin (137)
Working (54)
Nonworkin (284)
Working (100)
Nonworkin (329)
Working (111)
Nonworkin (338)
Working (116)
Nonworkin (249)
Working (82)
Nonworkin (240)
Working (81)
Nonworkin (358)
Working (119)
Nonworkin (234)
Working (87)
Nonworkin (367)
Working (122)
Nonworkin (244)
Working (72)
Nonworkin (302)
Working (99)
300
250
200
150
100
50
0
Nonworkin(377)
Working (123)
MEAN RANK
SIGNIFICANT DIFFERENCES AGE RECALL (WORKING AND NONWORKING MOTHERS)
Peace Interst Excite ShynesReflect
TABLE XXXIX SIGNIFICANT DIFFERENCES IN AGE RECALL AMONST WORKING
AND NON-WORKING MOTHERS
*Working mothers, who tend to belong to Higher Socioeconomic Groups and are also
generally better qualified, recall the ages at which they saw many emotions in their
children consistently earlier.
laugh
peace
reflect
working (114)
shy
nonworking (362)
working(117)
nonworking (357)
working(123)
nonworking (380)
working (121)
hope
nonworking (334)
guilt
nonworking (371)
working (118)
nonworking (357)
working (118)
shame
working(118)
sad
nonworking (352)
SIGNIFICANT DIFFERENCES IN ATTRIBUTIONS OF SOCIAL
SIGNIFICANCE: WORKING/NW MOTHERS
working (123)
350.00
300.00
250.00
200.00
150.00
100.00
50.00
.00
nonworking (380)
MEAN RANK
46
TABLE XXXX SIGNIFICANT DIFFERENCES IN ATTRIBUTIONS OF SOCIAL
SIGNIFICANCE AMONGST WORKING AND NON-WORKING MOTHERS
* Again, similar results obtain for the mothers when grouped for income and educational
qualifications. Working mothers felt sadness, hope, laughter, peace and reflection should be
encouraged in children. On the other hand, non-working mothers felt emotions like shame,
guilt and shyness should be encouraged. Except for sadness and laughter, the significance
level was very high for these eight emotions,.
100
90
80
70
60
50
40
Nonworking
30
Working
20
10
0
LOW
HIGH
Attended school
LOW
HIGH
Graduate
LOW
HIGH
Postgraduate &
abv
TABLE XXXXI
COMBINED SOCIO-ECONOMIC DEMOGRAPHIC OF THE MOTHERS IN OUR SAMPLE
* Combined Demographic Chart for Lower SES Groups and Higher SES Groups,
showing trends.
47
280.00
270.00
260.00
250.00
240.00
230.00
220.00
Married
(261)
Unmarried
(248)
Married
(261)
Fear
Unmarried
(249)
Jealous
Married
(261)
Unmarried
(248)
Excitemnt
TABLE XXXXII SIGNIFICANT DIFFERENCES IN EMOTION PICTURE
RECOGNITION TASK AMONGST MARRIED AND UNMARRIED INDIAN MALES
*Married men are better at recognizing jealousy and excitement than unmarried men
who recognize fear better!
G. FINALLY, A CROSS-CULTURAL DATABASE:
Cultural Differences: Some Dutch and Indian Data
This database is extremely rich and we have not completed our analysis, so results are
minimal as yet. Here we presented 40 Dutch and 40 Hindi speaking subjects with a list of 100
abstract nouns (50 emotion nouns and 50 non-emotion nouns = 8000 responses) and then
placed a restriction on their answers by asking them to give us the first verb, rather than just
the first word, that they associated with this noun.
Here, we found that, to begin with, the Hindi responses far exceeded the Dutch responses in
both variety and length often by a percentage increase of over one hundred percent. There
were also very interesting differences between the Dutch and Hindi responses in terms of the
embodied semantics of the emotion words. Our initial surmise for this huge difference is that
both the grammatical structure and the cultural context in which Hindi operates permit this
huge variation. Here is the list of words, in random order, that we gave respondents:
48
disappointment
interest
Comfort
Failure
Belief
Faith
Honesty
annoyance
panic
curiosity
fear
epiphany
Justice
Disturbance
Intelligence
anxiety
guilt
Romance
Sacrifice
Modernity
kindness
disgust
Dedication
shame
Talent
contempt
Confidence
suspicion
Liberty
terrorism
Motivation
frustration
Adventure
regret
Generosity
resentment
happiness
Wisdom
Bravery
Poverty
boredom
inspiration
Knowledge
Tragedy
Wealth
Irony
Enhancement
sadness
love
Compassion
Freedom
euphoria
doubt
Hope
enthusiasm
envy
calmness
Tolerance
anger
loneliness
Optimism
Idea
Elegance
Opportunity
hatred
Friendship
shyness
Success
surprise
Sophistication
awe
affection
pity
Wonder
grief
embarassment
desire
gratitude
pride
apathy
Humility
Reality
jealousy
Power
Trust
anticipation
hostility
Satisfaction
Evil
Imagination
humiliation
Improvement
Courage
confusion
Peace
Charity
Honour
Brutality
Rationality
Rumour
TABLE XXXXIII LISTS OF WORDS GIVEN FOR WORD ASSCOCIATION TASK
BY HINDI AND DUTCH SPEAKERS
Here is some indication of the variation in the Dutch and Hindi responses for the
emotion words:
satisfaction
resentment
shame
bravery
envy
hatred
surprised
pride
anger
boredom
humiliation
gratitude
disgust
regret
compassion
hope
panic
-50
Series1
0
50
100
150
200
TABLE XXXXIV: VARIATIONS IN DUTCH AND INDIAN RESPONSES
49
Main finding: Hindi responses are extraordinarily varied and individual compared to
Dutch responses which are far more semantically cohesive and show far greater withingroup consistency. The maximum ranges of Hindi responses over Dutch responses is
with the following words that show more than a 100% difference:
panic – 178%; happiness – 175%; sadness – 156% (basic emotions).
hope – 118%; and curiosity – 104% (non-basic emotions).
Out of a total of 51 emotion words analysed Dutch responses are greater in number
than the Hindi responses for only 4 words: guilt, satisfaction, suspicion and pity.
References to the body also appear to be far more specific and overt in the Hindi responses.
In the Hindi, we find from an analysis of 1600 responses so far that direct reference to body
parts in Hindi are mostly associated with emotion words.
TABLE XXXXV DIRECT REFERENCES TO THE BODY IN HINDI DATA
ANALYSING DUTCH DATA FOR THE SAME WORDS (SUMIT) + MLU
50
IV CONCLUSION: A SUMMARY OF TWENTY-ONE SIGNIFICANT FINDINGS/RESULTS
1. There is clear evidence of an order of emotional acquisition based on 505 mothers' recall
of the ages at which they saw particular emotions occurring in their children - from the basic
emotions beginning with happiness to the more complex 'moral' and self-conscious emotions
such as shame, guilt and reflection (24 emotions studied altogether). This is a robust
quantitative as well as qualitative result (Table VI in paper)
2. Emotions based on mother's memories appear to emerge in linked age-clusters. Based on
this clustering, seven well-defined stages of emotion acquisition may be postulated (Table
V). Further and importantly, a reading of this stage-by-stage chart reveals a balancing of the
order of display 'negative' and 'positive' emotions until children are nearly 3 years of age.
3. There is a striking parallel with the 'critical period' of language acquisition with the
emotions emerging over a three-to four year period in tandem with the growth of language.
In our data, both qualitative and quantitative analysis has shown this. (see especially, Table
XVIII, after Lennenberg)
4. Qualitative analysis of the narrative data supports the above findings and moreover has
given us a 'core vocabulary' associated with each of the 24 emotions studied (see charts and
power-point on 'modularity, narrative speech-acts emotional acquisition' (See Appendix IV).
5. Based on the careful 'picture-recognition' data-base that we have created, adults (reports by
1015 adults - 505 females and 510 males) that all adults can with high rates of success
identify at least 24 emotions with a very high degree of accuracy (85 to 90%, see Table I)
6. Females are significantly better than males at recognizing many emotions (Tables II
&III).
7. Mothers in joint families recall the age at which they see almost 75% of the 24 emotions
studied earlier that mothers in nuclear families (Table XXIX).
8. Mothers in higher socioeconomic groups recall the age at which see emotions in their
children earlier but mothers in lower SES groups are significantly better at identifying the
negative emotions in the picture recognition task. In particular, the basic emotion of fear is
51
recalled earlier and recognized better by mothers in lower SES groups who are also better at
recognizing sadness, boredom, jealousy and reflection. If mothers belong to the higher
socioeconomic group, they are significantly better at identifying pride, peace and disgust.
(Tables XXX and XXX1)
9. On emotions that were either encouraged or discouraged, there is general agreement in
our data with significant differences between the two groups of mothers from different SES
backgrounds emerging on only 6 out of the 24 emotions (Table XXXII). Here, less educated
mothers feel sadness and reflection (often interpreted as 'brooding') should be socially
discouraged while more educated mothers feel guilt and especially shyness should be socially
discouraged. With all the 24 emotions, graduate mothers stand in between the two poles in
their views. Even where there are significant differences, there is most convergence in
mothers' judgments of the expressions of interest and laughter (Table XXXVI) but here, too,
laughter and interest are discouraged to a significantly greater extent amongst lower SES
groups suggesting that a 'show' of emotion is not regarded as 'fitting' by these mothers and are
perceived as stigmatising or comprising a threats to others' face (Table XXXXII). This
finding emerges strongly, as well, from the narrative data (about 765 narratives analysed by
210 mothers using MAX QDA and some CA with fully transcribed data (Tables XIX and
XXV).
10. Neither the gender nor birth-order of the child significantly influence mothers' recall.
This surprising finding in the Indian context could be because the mothers in this study were
urban middle-class mothers and the family sizes were small with, on average, less than 2
children in every family.
11.
When mothers respond that they 'do not remember' or 'do not see' a particular,
overwhelming negative, emotion, the second category ('do not see') exceeds the first by
almost two or three hundred percent (Tables IX, X and XI). In other words, where the
emotion is not seen, it obviously be cannot be recalled. What’s fascinating is that in the case
of the basic emotions even when these are negative (fear, anger, sadness and disgust), the
mothers report seeing these and seeing them early (see Pie charts, Table XII).
12. Mothers' recall of the more complex emotions (contempt, pride, hate) presented a
different story, however. Here they were quite sure, in many cases, that their child had not
displayed the emotion in question. Contempt was the most dis-preferred of these emotions.
52
13. Disgust was the most accurately identified of all the 24 emotions across 1015 subjects
(Table I).
14. Contempt present an interesting case of a ‘self-conscious’ emotion that was lowest on
nearly every count: It was low on picture recognition index (71%); low on recall agreement
(fewest mothers, only 117 out of 505, only about one-fourth, recalled seeing this emotion in
their children); low on narrative embedding (only 15 out of 765 narratives devoted to this
emotion); low on frequency of repeated codes in narrative (brother, sister, parents, fight)
referring to this emotion'. It also had the lowest score amongst the ‘encouraged’ emotions. In
this respect it contrasted directly with happiness (see 15 below)
15. Happiness was the emotion recalled earliest by all the 505 mothers in our study. It was at
the other end of a socially abhorred emotion. It was high on recall agreement and accuracy,
very high on narrative embedding (accounting for 65 narratives out of 265); high on
frequencies of repeated codes (smiling, laughing etc.) and had the highest ratings by mothers
as an 'encouraged' emotion (see Narrative Graphs, Tables XIX to XXV). We have since done
a number of follow-up studies among, for example, about 600 IIT students on happiness
scores and find that we can now proceed with a more nuanced analysis of this and other
emotional parameters (student questionnaire and non-parametric analysis available on
request).
16. The most narrativised emotions, based on an analysis of 210 mothers' audio recordings
were happiness (65 narratives); anger (44); sadness (42); shame (40); surprise (36); anger
(35). The least narrativised emotions were: interest (23); reflection/deep thinking (22); hate
(22) and contempt (15) (see Narrative Graphs).
17. Narrative data analysis also shows that there although the evidence seems strong for the
basic emotions emerging much earlier in children, there is relatively weak evidence for
‘pure’ emotional states, since narratives involving, let’s say, a ‘basic’ emotion like anger
also refer to a range of other negative emotions such as irritation, disgust, shame etc. In other
words, there appear to be a clustering of emotional cognitions. We need to devise methods
to separate these possible ‘semantics-based’ conflations which are, of course, quite different
in different languages (see Narrative Graphs)
53
18. Among the males, married men are better at recognizing jealousy and excitement than
unmarried men who recognize fear better (Table XXXXII).
19. Culturally, a key-text that is still widely circulated today in India and powerfully
motivates the discourse and practices of aesthetics, especially dance, is Bharata's
Natyashastra based on a theory of nine representative emotions (shringara/rati: love; hasya:
humour (joy, mockery etc.); karuna/shoka: pathos, sorrow; raudra: anger, wrath; veera:
heroism; bhaya: fear, terror; bibhatsa: disgust; adbhuta: wonder; shanta: peace). What is
noteworthy is that Bharata's list of emotions tallies quite closely with Darwin's and, most
notably, all the putative basic emotions (happiness or joy, sadness or sorrow, anger, surprise,
fear and disgust) suggested by Darwin and later researchers are part of Bharata's list as well
(see Appendix 2).
20. Culturally again, in contrasting 40 Hindi and 40 Dutch speakers responses to 100
abstract and emotion noun-verb associations (a total of 8000 responses), we found that there
is a marked influence of the grammatical ‘explicator compound’ verbal structure on Hindi
subjects allowing them to dramatically vary both the length and semantics of their
associations, while Dutch speaker had a much smaller semantic scope and mean length of
utterance (MLU).
21. The words associated with the sixteen Darwinian emotions also varied greatly; for
example, many Dutch subjects associated ‘disgust’ (a basic emotion) with ‘vomiting’
whereas not of the Indian subjects did, associating it with ‘moving away’ instead, The
concepts ‘god’ and ‘prayer’ occurred semantically across the Indian data-set but in the Dutch
dataset only occurred in association with ‘faith’ etc. (other results are given in the paper
below) (see Graphs XXXXI1-XXXXV)
Overall, we believe we have obtained very robust publishable trends for the ages at which
certain emotions evolve as well as for the cultural parameters that might serve as
influences. It has not been possible to include all the results we have obtained – and are still
working on! – in this brief summary. However, we wish to emphasize that our study
constitutes one of the largest data-bases on emotional development and cognition and is
certainly the first of its kind in India.
54
In addition, we have sought to develop new non-invasive tools for the study of the emotions,
including the three stage recognition, recall and report methodology which spans both
qualitative and quantitative methods of further probing the acquisition of a putative EAD in
the human species, especially amongst children.
These initial studies now need to be augmented by careful experimental and laboratory
studies that will require collaboration with cognitive neuroscientists, for example, with those
studying ‘micro-expressions’ through eye-tracking instrumentation or dealing with
representations of hands, eyes and mouth in the motor-strip area of the human brain
(represented by ‘Penfield’s homunculus’ below). This stage of research is in the future and
here we would hope for collaborative experimental work with centres abroad as well as with
national centre such as NBRC and AIIMS, not to mention other departments within IITD.
V. CODA AND ACKNOWLEDGMENTS:
The analyses presented in this compendium are far from complete and there is lot more to be
done, especially in terms of theorizing what we have found so far. For the time being, though,
we would like to end by mentioning just one peripheral but startling insight that we got that
we had certainly not planned for but which came our way through serendipity.
This concerns the expression of emotions in animal mothers - something Darwin would have
appreciated. It happens that one of our film-makers was able to capture the unique maternal
conduct of a female monkey from the Himachal region of India. This monkey could not have
children of her own because the males in her pack had been neutered.
55
She therefore developed a unique and poignant strategy. She picked up and attempted to
nurse very young puppies, always protecting these 'babies' from attack. One of our filmmaker, Muneesh Sharma, captured the mother in action as she attempted to rear one such
baby. Tragically, her puppy babies eventually died each time as they could not adapt to her
‘monkey-care’ which included being fed grass and carried across roof-tops. This is extremely
rare footage which takes us, as mentioned, into the ‘Darwinian’ realm of animal maternal
emotions and constitutes a potential extension of our project.
Meanwhile, we would like to thank our film-makers for the splendid work they have done on
putting together the ethnographic database for our 'Darwinian Emotions': Arumugam,
Meenakshi Borooah, Nawed, Chandrika, Ajai Chawla, K.P. Madhu, Muneesh Sharma and
Kapil Sharma.
Further, we acknowledge the cooperation of the model for our 96-emotion stills database,
Ilina Bhaya-Grossman, who was a model in more ways than one.
The support of Prof. Veni Madhavan of the Indian Institute of Science (IIAS), Bangalore, Dr.
Falk Huettig of the Max Planck Institute (MPI), Nijmegen and Dr. H.B. Singh of the
Department of Science and Technology (DST), New Delhi on this project was valuable.
We also take this opportunity to express our gratitude to our International Advisory Board.
We look forward to their further support and advice as we consider possible extensions
of our initial project: Michael Bamberg (Clark University), Aniruddha Das (Columbia),
Susan Ervin-Tripp (Berkeley), Peter Gardenfors (Lund), Caroline Humphrey (Cambridge
University), Douglas Medin (Northwestern), Elinor Ochs (UC, Los Angeles), Roger Penrose
(Oxford University), Steven Pinker (Harvard University), VS Ramachandran (UCSD), Chris
Sinha (Portsmonth), Michael Toolan (Birmingham), Jef Verschueren (Antwerp).
Finally, we could not have gotten anywhere with our study without the generosity, honesty,
grace and perspicacity of the mothers and children from whom we learnt much.