enacting archetypes in movies: grounding the …...enacting archetypes in movies: grounding the...
TRANSCRIPT
This article was downloaded by: [Eindhoven Technical University]On: 05 July 2015, At: 14:58Publisher: RoutledgeInforma Ltd Registered in England and Wales Registered Number: 1072954 Registered office: 5Howick Place, London, SW1P 1WG
Click for updates
Digital CreativityPublication details, including instructions for authors and subscriptioninformation:http://www.tandfonline.com/loi/ndcr20
Enacting archetypes in movies: groundingthe unconscious mind in emotion-drivenmediaHuang-Ming Changab, Leonid Ivoninab, Marta Diazb, Andreu Catalab, WeiChena & Matthias Rauterberga
a Department of Industrial Design, Designed Intelligence Group, EindhovenUniversity of Technologyb CETpD Research Centre, Technical University of CataloniaPublished online: 13 Aug 2014.
To cite this article: Huang-Ming Chang, Leonid Ivonin, Marta Diaz, Andreu Catala, Wei Chen & MatthiasRauterberg (2015) Enacting archetypes in movies: grounding the unconscious mind in emotion-driven media,Digital Creativity, 26:2, 154-173, DOI: 10.1080/14626268.2014.939985
To link to this article: http://dx.doi.org/10.1080/14626268.2014.939985
PLEASE SCROLL DOWN FOR ARTICLE
Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”)contained in the publications on our platform. However, Taylor & Francis, our agents, and ourlicensors make no representations or warranties whatsoever as to the accuracy, completeness, orsuitability for any purpose of the Content. Any opinions and views expressed in this publicationare the opinions and views of the authors, and are not the views of or endorsed by Taylor &Francis. The accuracy of the Content should not be relied upon and should be independentlyverified with primary sources of information. Taylor and Francis shall not be liable for anylosses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilitieswhatsoever or howsoever caused arising directly or indirectly in connection with, in relation to orarising out of the use of the Content.
This article may be used for research, teaching, and private study purposes. Any substantialor systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, ordistribution in any form to anyone is expressly forbidden. Terms & Conditions of access and usecan be found at http://www.tandfonline.com/page/terms-and-conditions
Enacting archetypes inmovies: grounding theunconscious mind inemotion-driven mediaHuang-Ming Changa,b, Leonid Ivonina,b,Marta Diazb, Andreu Catalab, Wei Chena
and Matthias Rauterberga
aDepartment of Industrial Design, Designed Intelligence Group, EindhovenUniversity of Technology; bCETpD Research Centre, Technical Universityof Catalonia
[email protected]; [email protected]; [email protected]; [email protected];[email protected]; [email protected]
Abstract
The concept of emotion-driven media integrates affec-tive computing into developing new types of digitalmedia by including the human spectator as an essentialpart of the whole system. Enactive media, as a landmarkin this direction, provides a theoretical basis that isrooted in the enactivist approach in cognitive science.Based on this thesis, it is required to establish a practicalprocedure for developing the media content throughintegrating psychological approaches. We demonstratea systematic approach for developing the content foremotion-driven media. The theory of archetypes, oneof the psychoanalytical approaches, was applied forediting movie clips as the media content, and coupledthe content with emotional responses by using exper-imental psychological methodologies. The resultsshowed that it is promising to utilise archetypal movieclips as the content of emotion-driven media and arche-typal symbolism can be a useful resource for developingthe content of media systems.
Keywords: enactive media, emotion, archetypes,movie, psychological approach
1 Introduction
In the field of human–computer interaction(HCI), there have been tremendous endeavoursin recent years to build an intelligent systemthat is capable of communicating with humansin a natural way. Affective computing, as abranch of HCI, specifically aims at enhancingthe communication between human andmachine by expanding the emotionally chargedchannel (Scheirer and Picard 1999). Researchershave made considerable progress moving fromuser interfaces based on the physical level ofuser activity to a multimodal interaction thattakes into account affective states of users. Thishas expanded the vision of research in the HCIfield, and pushed HCI into the next paradigmwhere interaction goes beyond the traditionalinput–output mechanical process through aphysical interface. Various applications of phys-iological technology are employed to measureand recognise human emotion, so that intelligentsystems are empowered to react not only tousers’ explicit conscious behaviour but also tousers’ unconscious affective states. The advancesin affective computing considerably facilitate the
Digital Creativity, 2015Vol. 26, No. 2, 154–173, http://dx.doi.org/10.1080/14626268.2014.939985
# 2014 Taylor & Francis
Dow
nloa
ded
by [
Ein
dhov
en T
echn
ical
Uni
vers
ity]
at 1
4:58
05
July
201
5
development of applications in different fields,especially digital media.
Tikka (2010) has set a nice example of newparadigms for applying affective computing indeveloping digital media: enactive media. Thetheoretical foundation of the idea originates fromenactivism in cognitive science, which takes a‘groundless’ stance that resists the mind-bodydualism, and argues that mind, body and worldare interrelated and interdependent (Varela,Thompson, and Rosch 1992). Extending enacti-vism to the cinema experience, Tikka considerscinema as a metaphorical externalisation of embo-died mind, saying that the spectator’s mind and theimage of the cinema emerge and interplay withand within each other (Tikka 2004, 2006). Thatis to say, the spectator and the cinema have to beput together into a systematic context thatmutually and continuously creates its ownmeaning. Based on this phenomenological argu-ment, she made an attempt to develop a mediasystem that technically brings the spectator andthe narratives together in an enaction (Tikka2010). Two of the most important notions of theenactive mind theory are autonomy and sense-making (Froese and Di Paolo 2011). The idea isthat a living organism (i.e. the spectator) is auton-omous in a way that it not only responds to theexternal world (i.e. the media) in the traditional
sense of producing the appropriate action for agiven situation, but actively regulates the con-ditions of its exchange with the world (Di Paolo,Rohde, and De Jaegher 2010). This exchangeprocess is inherently significant to the livingorganism, who directly participates in the gener-ation of meaning by his or her action. In essence,the spectator enacts the whole system; thesystem would be disembodied without any specta-tor; an enactive media system together with itsspectator creates a looping cycle to form a self-contained and self-organising system—a systemthat the autopoiesis theory described as a livingorganism (Maturana and Varela 1980). To enablethe whole system, emotion, as the key couplingbetween the spectator and the media, is highlightedas the central part of the functioning mechanism(see Figure 1). The content of the media systemdynamically changes with the spectator’s real-time emotional states while her emotional statesare also influenced by the generated content(Tikka 2010). The concept of enactive media notonly establishes a new paradigm of emotion-driven media systems, but also raises many inter-esting research questions on various aspects. Forexample, how does the designer of the mediasystem develop the content? And how does thedesigner couple the content and the spectator’semotional states in a methodologically valid way?
Figure 1. The conceptual framework of an emotion-driven media system.
Grounding the unconscious mind in emotion-driven media
155
Dig
italC
rea
tivity,V
ol.
26
,N
o.
2
Dow
nloa
ded
by [
Ein
dhov
en T
echn
ical
Uni
vers
ity]
at 1
4:58
05
July
201
5
Emotion is assumed as the mediator betweentwo parallel processes: consciousness and uncon-sciousness (Rauterberg 2010). While consciousthought and intentional behaviour have beeninvestigated for a long time, the importance ofthe unconscious mind has long been overlookedby mainstream science and just started to attractmore attention to contemporary research (Westen1999). The common biased view in cognitive psy-chology often equates unconsciousness with sub-liminal information processing, which is tootrivial to enter our consciousness. However, agrowing body of recent studies has indicated thatmany thoughts, behaviours and decisions areformed or made unconsciously before we are con-sciously aware of them (Dijksterhuis 2004; Giger-enzer 2007; Wilson and Bar-Anan 2008; Custersand Aarts 2010). It is suggested that the uncon-sciousness functions as a higher level of humanmind that precedes the arrival of consciousnessin terms of its unintentional nature and the inherentlack of awareness (Bargh and Morsella 2008).While the focus of mainstream science attends tothis new direction, the field of psychoanalysisand psychotherapy has long been studying theunconscious mind. Psychologist and psychiatristCarl Jung proposed the theory of archetypes, ela-borating a universal inherent tendency in allhumans’ unconscious mind (Jung 1959). Thistheory grounded the basis of analytical psychol-ogy and has been used to explain other psycho-logical concepts (e.g. Percival 1993; Engel 1984;Martin, Carminati, and Carminati 2013). Thetheory of archetypes has influenced various disci-plines, e.g. research in personality (Munteanu et al.2010; Hautala and Routamaa 2008; Robertson2013) and marketing research (Tsai 2006; Lim,Acito, and Rusetski 2006; McLoone 2010; Wood-side, Megehee, and Sood 2012; Caldwell, Henry,and Alman 2010). One of the extended theoriesof archetypes—archetypal symbolism—was pro-posed to document different forms of archetypalcontent that were distributed in different culturesand religions all over the world (Ronnberg andMartin 2010). It became an important resourcefor research on media and movies (Faber andMayer 2009; Hauke and Alister 2001). However,
very few studies have attempted to investigatethe emotional responses to archetypal content. Itappears that archetypal symbolism could be apromising resource for developing content forinteractive media.
Although media content can be presented invarious forms, e.g. pre-recorded film clips, audiotracks or text excerpts, or even real-time beha-viours (Kaipainen et al. 2011), the narrative ofmedia content plays the key role in delivering theartistic meaning of the whole work to its spectators.For developing emotion-driven media systems, themost challenging part is to make the content mean-ingful in terms of emotions. Designers need to takeinto account what emotion is to be induced while aspectator is watching a specific scene, and thenconsider how this induced emotion can connectto other emotions to form a story. Most researchershave used a story-based approach. They positionthe spectator as a second author to tell theauthor’s story in different ways, i.e. the second-order authorship (Tikka 2010). The designerneeds a complete story as the basis for the mediacontent, and then decomposes its storyline into anumber of meaningful scenes. By mappingproper emotional states to each scene, the authorcould construct a narrative space as a predefinedscript for dynamic storytelling (Tikka, Vuori, andKaipainen 2006). In other words, the originalstory remains, but the storyline is dynamicallyreshuffled by the spectator according to the real-time changes in his or her emotional states.However, this approach is specific for one storyand cannot be generalised to other stories. Wepropose a new approach—the theme-basedapproach—to prepare the content without takingan initial story as a reference. This approachrequires meaningful ‘themes’ to be a generic classi-fication for specific scenes and thus collect scenesacross different movies according to these themes.In order to apply these themes in emotion-drivenmedia systems, the links between the themes andcorresponding emotional responses have to beunique. It is necessary to examine whether thecontent in each theme induces a unique pattern ofemotional responses that can be reliably recognisedby computational systems. The theme-based
Chang et al.
156
Dig
italC
rea
tivity
,V
ol.
26
,N
o.
2
Dow
nloa
ded
by [
Ein
dhov
en T
echn
ical
Uni
vers
ity]
at 1
4:58
05
July
201
5
approach allows emotion-driven media to providemixed content that is related to similar experiences.In terms of authorship, the spectator becomes thefirst author who creates a new story by mixingvarious movie clips in the same theme.
In the present study, we applied archetypalsymbolism as the main reference for classifyingscenes in different movies. In order to verify thevalidity of this classification, we developedseveral sets of movie clips and conducted anexperiment to examine whether these archetypalmovie clips are suitable for emotion-drivenmedia systems. The developing process consistsof two stages: edit and verification. We first uti-lised a Jungian approach to collect seven cat-egories of archetypal clips edited fromcommercial movies that are available in themarket; the categorisation follows the theory ofarchetypes. Next, in order to verify the feasibilityof using our archetypal movie clips as thecontent of emotion-driven media systems, it isnecessary to evaluate whether archetypal movieclips in different categories can be differentiatedbased on the participants’ emotional responses.We conducted a study that applies a typicaldesign in psychological experiments specificallyfor studying emotion. In this experiment, weharnessed two different techniques for emotionrecognition—self-reports and physiologicalmeasures—to capture conscious and unconsciousemotional responses simultaneously. This allowsus to further discuss whether it is possible for thespectators to interact with the archetypal contentof the media system with their unconscious mind.
2 The collective unconscious andarchetypes
After the idea of the unconscious mind was popu-larised by Sigmund Freud, Carl Jung (1959)proposed a broader concept: the collective uncon-sciousness, based on his own observation of hispatients during their therapeutic sessions and hisextensive research on myths, religions and cul-tures. This theory suggests that all human beingsshare a deep level of unconscious mind univer-sally, which is hidden beneath the personal
psyche (including personal conscious and per-sonal unconscious minds). The collective uncon-sciousness contains some contents and modes ofbehaviours that are identical in all human beings,and thus constitutes a common psychic substrateof a universal nature that is present in every indi-vidual. As a result, all human beings, sharingessentially the same biological equipment (e.g.the central nervous system), would show a ten-dency to perceive common meanings embodiedin a symbol, even at an unconscious level(Bradshaw and Storm 2013).
Jung further proposed the concept of arche-types (Jung 1964). Archetypes are defined as thecomponents of the collective unconsciousness,which is an inborn tendency that cannot be con-sciously acquired to experience things in acertain way. They exist universally in the psyche,and prepare individuals psychologically to dealwith life experiences that are universallycommon (Walters 1994). Archetypes are uncon-scious psychic impulses that are impersonal, inher-ited traits that present and motivate thoughts,emotions and behaviours long before any con-sciousness develops. An extreme analogy wouldbe describing archetypes as the structure of thepsyche, which is similar to organs of the physicalbody (Jacobi 1973). In this way, archetypes aresimilar to other sensory and cognitive models,e.g. receptive fields of the retina are not con-sciously perceived, but determine the structure ofvisual perception (Perlovsky 2007). Someresearchers consider archetypes as a universalknowledge that is rooted in our unconsciousness(Chang et al. 2013a). Humans interpret the worldwith archetypes, though they are unaware of theexistence of these archetypes. These theorieshave built a basic framework of the unconsciousmind, and further became the theoretical basis ofanalytical psychology and Jungian psychotherapy(Roesler 2012; Knox 2001).
Since archetypes are in the deepest level ofunconsciousness, it was claimed that archetypescannot be directly described in a conscious wayor in a logical sense, but archetypes can beobserved indirectly through symbols (Jung1964). Archetypes can be manifested by ancient
Grounding the unconscious mind in emotion-driven media
157
Dig
italC
rea
tivity,V
ol.
26
,N
o.
2
Dow
nloa
ded
by [
Ein
dhov
en T
echn
ical
Uni
vers
ity]
at 1
4:58
05
July
201
5
people as observable patterns such as repeatedthoughts, ideas or imaginings throughout historythat come to be represented in symbols foundglobally in artwork, religion and myth (Bradshawand Storm 2013). These phenomena are calledarchetypal images—the symbolic representationsof the activation of archetypes that are manifestedin infinitely many forms of content (Jones 2003).Although its form might vary, the archetypeitself always remains intact (Nunn 1998). Inaddition to ancient artwork and myths, manyarchetypal images are still manifested in modernsociety in many different ways, especially inmass media. For decades, Jungian scholars havekept documenting the archetypal images inancient cultures and exploring archetypal imagesin modern society (Gronning, Sohl, and Singer2007). They founded a field of study, ‘archetypalsymbolism’, to investigate the essence of howthe world is originally conceived through spon-taneously interpreting the symbolic meaning ofall the phenomena that emerge in a human mind.Movies, as a contemporary form of creativeartwork that reflects societal and cultural phenom-ena, appear to be a fruitful resource for research onarchetypal images. Several researchers have usedthe theory of archetypes for movie analysis(Cabot 2011; Heyraud 1998; Broda 1994;Zehnder and Calvert 2004; Hauke and Alister2001). From a Jungian standpoint, since allhuman beings share archetypes in their collectiveunconsciousness, all these archetypal imagesexisting in the media that are manifested byhumans conversely shape our unconsciousthoughts and behaviours.
Jung’s theories were often criticised by themainstream cognitive science for the unfalsifiabil-ity of his explanation about the collective uncon-sciousness (Walters 1994; Jones 2003). Althoughthe theory of archetypes has not yet been con-firmed in terms of empirical evidence, it isbroadly used for analysing narratives, such asmyths (Campbell 1973) and movies (Hauke andAlister 2001). In the context of media design, itappears that archetypal symbolism can be posi-tioned as a design pattern for creating symboliccontent, such as narratives. When it comes to
emotion-driven media systems, it is necessary todetermine whether this design pattern can becoupled with specific emotional responses acrosspeople in order to serve as a valid reference fordesigners to create new content. Our goal wasnot to justify the underlying hypothesis of thetheory of archetypes, but to evaluate the feasibilityof applying this theory to developing mediacontent.
3 Archetypes in movies
Movies are a very complex form of symboliccontent that communicates delicate and richvisual-audio information to the audiences bymeans of storytelling. Due to its special contextualsetting of viewing a movie in a cinema, it offers notonly visual-audio narratives, but also both a meansand a space similar to psychotherapeutic sessionsthat enable the viewer to witness their psyche inprojection (Hauke and Alister 2001), whichmeans that spectators are in a process of projectingthe image of themselves in the movies and not justbeing an indifferent observer. Moreover, it is alsoclaimed that the cinema delivers a contemporaryexperience for the spectators to set apart fromtheir physical life and engage their unconsciousin a manner similar to hypnosis and dreaming.This might resonate to the concept that Tikka(2004, 2006) intended to put forward: cinema asa model of mind. As Tikka suggests, the cinemaexperience itself is already an implementation ofthe enactive mind. Since archetype is unconsciousknowledge that cannot be directly accessed, a phe-nomenological stance has to be taken in order toreveal the symbolic meaning of archetypalimages in movies.
Some researchers have put efforts into elabor-ating symbolic meanings on modern moviesaccording to Jung’s theory (Cabot 2011;Heyraud 1998; Broda 1994; Zehnder and Calvert2004; Hauke and Alister 2001). These effortshave provided good references for our analysisof archetypes. A typical Jungian approach to ana-lysing creative works is suggested to take a vision-ary standpoint (Jung 1967). Instead of viewing thewhole work of art as a physical entity to decom-
Chang et al.
158
Dig
italC
rea
tivity
,V
ol.
26
,N
o.
2
Dow
nloa
ded
by [
Ein
dhov
en T
echn
ical
Uni
vers
ity]
at 1
4:58
05
July
201
5
pose, the analyst should take a first-person point ofview of the character in the scene to actively gothrough the situation, and describe the experiencethat is relevant to archetypes. It is similar to aprocess of embodied simulation (Tikka 2010).The analyst should experience the fictional worldwith her own experience, taking into account thecontext of the scene, other figures that the charac-ter is interacting with, and the mental states and thepersonality of the character that is being simulated.Rather than logical reasoning, the Jungianapproach places more emphasis on associatingand experiencing. By living in the moment thatthe character is experiencing, the symbolicmeaning of archetypes would emerge onlythrough the enaction between the analyst and themovie. It appears that the enactive approach andthe Jungian approach fit well together within thecontext of the cinema experience. While the enac-tive approach is more sophisticated in philosophi-cal discourses of the human mind, the Jungianapproach focuses more specifically on the sym-bolic meaning of the archetypes in movies.
Jung first discovered several essential arche-types: hero; shadow; anima; animus; mentor(or wise old man); and mother (Jung 1969;Neumann 1970). Furthermore, mythologistJoseph Campbell (Campbell 1973) extendedJung’s concept of the hero archetype to a more com-plete scope—monomyth1 (or hero’s journey)—acommon structure in all mythical hero stories indifferent cultures and religions. This structuredepicts seventeen stages that the hero has to gothrough to complete his adventure. These stagescan be generally divided into four larger stages:departure; trials; rebirth; and return. In total, thereare nine archetypal categories that are consideredfundamental: anima; animus; four stages of thehero’s journey; mentor; mother; and shadow. Thehero archetype, as a symbolic form of the self, rep-resents a metaphorical instance of who lives his lifeas going through the hero’s journey repeatedly.Other archetypes, such as mentor, mother, etc., rep-resent the essential figures that a person would meeton his or her lifelong journey.
As one of the motivations of this study, wedeveloped a standard procedure for editing arche-
typal movie clips for the content of emotion-drivenmedia systems (see Figure 2). We applied this pro-cedure to analyse the abovementioned nine arche-types in movies and edited proper clips as thecontent for media systems. First, we surveyedmany commercial movies in the market to findsuitable movies that might contain narratives rel-evant to these target archetypes. Then, we selectedscenes that convey experiences of specific arche-types from a first-person point of view. In orderto edit movie clips that can be used in experimentsas well as in a media system, the format of the clipsneeds to be defined first. All the selected scenesshould be edited into the same format and adataset of movie clips built up that is ready forlater experiments. The format of the movie clipswe edited is described as follows:
. for adequate statistical power, at least three clipsin each category are required;
. the unified length2 of each clip is 60 + 6seconds;
. every clip contains scenes selected from onlyone movie;
. the transition between scenes uses fade-in andfade-out effects;
. the screen resolution of the movie clips is720×480 pixels with AVI file format.
To further confirm that the selected scenes authen-tically represent archetypes, we cooperated withJungian scholars from The Archive for Researchin Archetypal Symbolism (ARAS)3 to review thefilm clips we had collected. They did the reviewseparately for all the clips. Only those clips thatthey all agreed to consider as archetypal clipswere kept in our collection. After several iterationsof discussions and strict filtering sessions, we builta set of movie clips that are verified as archetypalimages. Although we attempted to use clips for allthe nine essential archetypes, we could not collectenough clips for animus and mother within thelimited time span (only two clips collected). Themain reason is that these two archetypes are lessexplicit, so it is extremely difficult to demonstratea full archetypal experience within one minute.Therefore, we proceeded with the seven categoriesof movie clips to the later experiment, which are:
Grounding the unconscious mind in emotion-driven media
159
Dig
italC
rea
tivity,V
ol.
26
,N
o.
2
Dow
nloa
ded
by [
Ein
dhov
en T
echn
ical
Uni
vers
ity]
at 1
4:58
05
July
201
5
anima; hero’s departure; hero’s trials; hero’srebirth; hero’s return; shadow; and mentor (seeTable 1). Full descriptions of the symbolic mean-ings of each archetype are reported elsewhere(see Chang et al. 2013b).
4 Coupling emotional responses withthe content
When an archetype becomes activated and isexperienced with associated feelings and thoughts,it will result in a complex within the personalunconscious (Walters 1994). According to Jung,a complex within the personal unconscious is anindependently organised conglomeration ofemotions and ideas that are specific to an individ-ual and are products of interactions among anumber of archetypes (Jung 1959). Although thestimulated emotions and thoughts cannot directly
be accessed by our conscious introspection, thephysiological measures still provide us withchances to infer the unconscious emotion. Sinceemotion plays a central role in coordinatinghumans and the media in an emotion-drivensystem, it is necessary to investigate the corre-lation between the elements of the content andthe induced emotional responses in order tobuild the interactive mechanism afterwards. Eachelement of the content should be indexed by corre-sponding emotional responses. However, the realchallenge is to confirm the validity of thecorrelations between emotional responses andthe content. Precisely recognising humanemotion from these physiological data is still chal-lenging (Picard 2003). This might be due to thefact that an ultimate definition of emotion is stillunavailable. This is not surprising, because forthe last century the same situation has also per-
Figure 2. The editing process for archetypal movie clips.Note: ARAS means The Archive for Research in Archetypal Symbolism.
Chang et al.
160
Dig
italC
rea
tivity
,V
ol.
26
,N
o.
2
Dow
nloa
ded
by [
Ein
dhov
en T
echn
ical
Uni
vers
ity]
at 1
4:58
05
July
201
5
Category: description Movie Start End
Anima:A female image that makes a male touchy, irritable,
moody, jealous, vain, unadjusted, and discontent.
American Beauty (Mendes 1999) 0:16:15 0:17:17
Malena (Tornatore 2000) 0:19:18 0:20:20
Perfume: The Story of a Murderer(Tykwer 2006)
0:18:03 0:18:180:21:20 0:22:15
Hero departure:A sudden tragedy happens to a young kid’s peaceful
life, so that he is forced to leave home.
V for Vendetta (McTeigue 2005) 0:41:55 0:43:03
Braveheart (Gibson 1995) 0:10:10 0:10:460:14:13 0:14:43
The Lord of the Rings: TheFellowship of the Ring (Jackson 2001)
2:21:122:22:372:23:10
2:21:472:23:062:23:16
Hero trials:A road of trials that one young hero has to go through.
He is mentally and physically devastated.
V for Vendetta (McTeigue 2005) 1:23:29 1:24:34
Braveheart (Gibson 1995) 2:07:392:08:47
2:08:372:08:58
The Matrix (Wachowski and Wachowski 1999) 2:02:25 2:03:25
Hero rebirth:A miracle happens. The wounded hero resurrects from
the fire and water and gains great power as a processof transcendence.
V for Vendetta (McTeigue 2005) 1:24:59 1:26:00
Braveheart (Gibson 1995) 2:15:392:17:35
2:16:152:18:01
The Matrix (Wachowski and Wachowski 1999) 2:04:35 2:05:45
Hero return:The hero returns home with the given great power.
But he has to give away everything, even his life,to achieve a higher value for his people. His spiritremains alive as a legend.
V for Vendetta (McTeigue 2005) 2:02:402:03:22
2:03:042:04:06
Braveheart (Gibson 1995) 2:48:562:49:112:49:54
2:49:082:49:532:50:09
The Matrix Revolutions (Wachowski andWachowski 2003)
1:53:401:54:021:54:331:55:241:56:02
1:53:471:54:051:54:501:55:391:56:29
Mentor:A wise old man who supports the hero along the way to
the early trials. He represents the support, theknowledge and the guidance that the hero needs whengrowing strong.
The Lord of the Rings: The Fellowship of the Ring(Jackson 2001)
2:03:05 2:04:10
The King’s Speech (Hooper 2010) 1:42:131:42:581:45:33
1:42:441:43:181:45:52
The Lion King (Allers and Minkoff 1994) 0:24:380:25:29
0:25:050:26:06
Shadow:The dark side of the self, personifies everything that a
person refuses to know about himself, representsinner desire, hate, and brings the deepest fear.
The Lord of the Rings: The Two Towers (Jackson2002)
1:35:19 1:36:20
Fight Club (Fincher 1999) 1:48:24 1:49:32
The Dark Knight (Nolan 2008) 1:24:22 1:25:30
Table 1. The descriptions for the categories of archetypes and the sources of the movie clips.Notes: The clips consist of one or more fragments that were cut from the movies at the times specified in the last two columns. Thetime format is [hours:minutes:seconds].
Grounding the unconscious mind in emotion-driven media
161
Dig
italC
rea
tivity,V
ol.
26
,N
o.
2
Dow
nloa
ded
by [
Ein
dhov
en T
echn
ical
Uni
vers
ity]
at 1
4:58
05
July
201
5
plexed psychologists, who are still debating thenature of human emotion (Lindquist et al. 2013).Some researchers claim that there are only asmall number of basic emotions that are ‘hard-wired’ into brain circuits, such as anger orsadness (e.g. Ekman and Cordaro 2011). Othersargue that emotions are constantly constructedby the human mind so that emotions should notbe of discrete, natural kinds (e.g. Barrett 2011).In the case of our study, it is contradictory totake the former assumption because the fundamen-tal assumption of the enactive mind is that psycho-logical events are enacted, not pre-existing. Itwould be more feasible to assume that emotionsare not just basic kinds but an ongoing psychologi-cal phenomenon that is enacted when a livinghuman is embodied in the world.
In light of this, we do not attempt to identifybasic emotional qualities in each clip due to thefundamental conflict of theory, but concentrateon whether these categories of archetypal movieclips can be differentiated by the emotionalresponses so that the media system is able toreact accordingly. To measure emotionalresponses in laboratory settings, researchersoften apply either self-report techniques or physio-logical measurement. The most simple andstraightforward way to measure the content of amental representation of emotion is often con-sidered to be self-reports, which rely on the partici-pant’s conscious introspection (Barrett et al.2007). However, this technique requires extraefforts from the participants that might distractthem from the affective stimuli (Soleymani et al.2012). On the other hand, physiological measure-ment directly monitors bodily changes in physio-logical states. These spontaneous bodily changesare initiated by the autonomic nervous system(Kreibig 2010), which provides a great opportu-nity to measure unconscious emotion (Miller1992). Since most of the previous studies in thisarea focus on conscious emotions, their exper-iments assumed that self-report results are thefact, that humans can honestly and precisely intro-spect about their current emotional state and reportit. Therefore, their goal is to examine how well thephysiological signal can explain these predefined
basic emotions (e.g. Rottenberg, Ray, and Gross2007; Soleymani, Pantic, and Pun 2012).
Since the emotion toward archetypal images isa new research topic, it is still unknown whether itis conscious or unconscious. Therefore, in order tohave a more comprehensive view of the emotionalresponses to archetypal images, it is suggested thatboth self-reports and physiological measures areincluded for emotion recognition (Chang et al.2013c). By comparing the results from the twomeasurements, it is possible to infer whether anyunconscious emotions are induced. Our firstresearch question is to ask whether each categoryof archetypal content induces unique emotionalresponses that can be classified with recognitionrates higher than chance level. If the answer tothe first research question is positive, the secondresearch question is to examine which techniqueof the emotion recognition (self-reports or physio-logical measures) is more robust for classifyingeach movie clip category.
5 Experiment
An experiment was conducted to examine whetherall the categories of archetypal movie clips can bedifferentiated from each other according to partici-pants’ emotional responses. For experimentaldesign, we followed the methodological paradigmin experimental psychology that has been welldeveloped specifically for using film clips to elicitemotions in laboratory settings (e.g. Rottenberg,Ray, and Gross 2007; Soleymani, Pantic, and Pun2012). The only difference is that our movie clipswere categorised based on archetypes instead ofbasic emotions. As these studies suggest, ourexperiment applied a within-subject design. Everysession accommodated one participant, and everyparticipant viewed all the twenty-one movie clipswe collected. Before the presentation of eachmovie clip, the participant was asked to adjusttheir breathing to regain the baseline emotionalstate in order to avoid the overlap of the emotionalresponses to the previous movie clip. As a double-blind design, movie clips were played in a randomorder. Neither the experimenter nor the participantknew the sequence of the playlist.
Chang et al.
162
Dig
italC
rea
tivity
,V
ol.
26
,N
o.
2
Dow
nloa
ded
by [
Ein
dhov
en T
echn
ical
Uni
vers
ity]
at 1
4:58
05
July
201
5
5.1 ApparatusThe experiment was held in the Ibux-Lab researchlaboratory at the CETpD Research Centre, whichwas arranged as a living room to make the partici-pant feel comfortable and relaxed sitting on a sofa.Movie clips were projected onto a white wall(325 cm by 175 cm). The experimenter couldobserve from outside of the room through a one-way mirror so that the participant was left alonewhile viewing the movie clips and providingself-report data. For gathering self-report data,the self-assessment manikin (SAM) (Bradley andLang 1994) was adopted. SAM is widely used inpsychological experiments (e.g. Soleymani et al.2012; van den Broek 2013). It applies the dimen-sional model to represent human emotion, usingthree dimensions to construct affective space:valence; arousal; and dominance. It is claimedthat these three scales are capable of representingmost of the commonly known human emotions.For physiological measurement, electrocardiogra-phy (ECG) and skin conductance data wererecorded with ShimmerTM wearable wirelesssensors (Burns et al. 2010) and sent to anAndroid device via BluetoothTM protocol.
5.2 ParticipantsTwenty-two volunteers participated in our study.Most of the participants were undergraduate orgraduate students at the Technical University ofCatalonia. We also recruited several participants
of an older age. Of the twenty-two participants,nine were women and thirteen were men. Theaverage age for the women was 28.67 years (SD¼ 9.39) and for the men it was 27.77 years (SD¼ 6.13). The participants had diverse nationalbackgrounds (four from Asia, fourteen fromEurope and four from South America). Werequired the participants to have normal or cor-rected to normal vision and hearing. Prior to theexperiment, each participant signed an informedconsent form and was later rewarded with a smallpresent for participation in the laboratory session,which took approximately 1.5 hours. The trialswere carried out according to the experimentalplan and all the collected data were analysed.
5.3 ProcedureThe whole procedure is demonstrated in Figure 3.The participant was invited to sit on a couch in thetest room at the laboratory. The participant wasthen asked to read and sign an informed consentform. After signing the agreement, the participantput on electrodes and physiological sensors fol-lowing the experimenter’s instructions. The con-nection of the sensors was checked while theparticipant was filling in a short questionnaire ondemographic data. Once the questionnaire wasfilled in, the experimenter gave a tutorial aboutthe experiment and an example of what the partici-pant should do throughout the session.
Figure 3. The procedure of the experiment.Note: Movie clips are played in a random order.
Grounding the unconscious mind in emotion-driven media
163
Dig
italC
rea
tivity,V
ol.
26
,N
o.
2
Dow
nloa
ded
by [
Ein
dhov
en T
echn
ical
Uni
vers
ity]
at 1
4:58
05
July
201
5
For the example test, a neutral clip extracted fromthe movie Coral Sea Dreaming: Awaken (Hannan2010) was used. The participant was taught how toprovide self-report data about her emotion byfilling out the SAM scale. We did not disclose anyinformation about archetypes or the content of themovie clips to the participant during the session.After the example test, the light in the laboratorywas dimmed to make the viewing experiencesimilar to a real cinema. Then the presentation ofmovie clips started. Before each clip, the participantfollowed a breathing-guiding video for 20 seconds toadjust the respiration rate of the participant to theinitial baseline. The physiological data recordedduring the video with a breathing pattern were laterused in the analysis as physiological baseline. Afterviewing each film clip, the participant provided a ret-rospective self-report by rating her emotion along thedimensions of the SAM with paper and a pen. Aftergiving ratings, the breathing clip would be shownagain to start the next presentation. The experimentended when the participant finished viewing theentire collection of twenty-one movie clips.
5.4 ResultsIn our experiment, each participant viewed all thearchetypal movie clips in a random order. Thus,the study had a repeated-measures design wherephysiological measurements were made on thesame individual under changing experimental con-ditions. Moreover, the participants providedreports via the SAM ratings after every experimen-tal condition. The multivariate analysis of variance(MANOVA) for repeated measures was applied toanalyse data and explore the relationship between
the categories of archetypal movie clips (indepen-dent variable) and the participants’ emotions(dependent variable) measured by SAM and phys-iological responses. The main effect of MANOVAtested whether the patterns of the participants’physiological responses were different betweenvarious categories. All statistical tests used a0.05 significance level.
The descriptive statistical analysis on the SAMdata is summarised in Table 2. The MANOVA forrepeated measurements was performed for theSAM ratings of valence, arousal and dominance.It demonstrated a significant main effect of thearchetypes presented in the film clips on theSAM ratings [F(18, 351.210) ¼ 10.060, p ,
0.001 (Wilks’ lambda)]. The results of theMANOVA conducted for the features extractedfrom the physiological signals indicated thatthere is also a significant main effect of the arche-types pictured in the film clips on the participants’physiological responses [F(216, 548.182) ¼1.310, p ¼ 0.007 (Wilks’ lambda)], which con-firms the significant relationships between the cat-egories of the movie clips and the SAM ratings aswell as between the movie clip categories and thephysiological responses. In order to examinewhether the emotional responses were universalacross different cultural backgrounds, we put‘countries’ as a between-subject factor in the stat-istical analyses. However, the sample size of thecurrent study is insufficient for performingbetween-subject multivariate tests (twenty-twoparticipants vs thirty-six features extracted fromphysiological signals) and the numbers of partici-pants from different countries were not balanced.
Archetype
Valence Arousal Dominance
Mean SE Mean SE Mean SE
Anima 5.879 0.291 4.515 0.429 5.803 0.353Hero departure 4.015 0.229 4.485 0.364 4.515 0.393Hero trials 3.864 0.266 4.909 0.388 4.439 0.345Hero rebirth 5.924 0.278 4.455 0.356 6.197 0.298Hero return 6.318 0.291 4.818 0.373 6.742 0.328Mentor 6.273 0.193 3.455 0.273 6.348 0.295Shadow 4.591 0.231 4.576 0.420 5.212 0.324
Table 2. Mean values and standard errors (SE) of the self-report results recorded by SAM ratings, ranging from 1 to 9.
Chang et al.
164
Dig
italC
rea
tivity
,V
ol.
26
,N
o.
2
Dow
nloa
ded
by [
Ein
dhov
en T
echn
ical
Uni
vers
ity]
at 1
4:58
05
July
201
5
We therefore looked into the results of analyses onthe interaction between categories of the clips andthe nationalities of the participants. For the SAMratings, there was no significant effect on thismatter [F(216, 156.887) ¼ 0.905, p ¼ 0.752(Wilks’ lambda)]. Similar results were obtainedfrom the analysis on physiological signals[F(2592, 1008.114) ¼ 1.091, p ¼ 0.051 (Wilks’lambda)]. The results suggested that cultural back-grounds did not introduce significant effects on theparticipants’ emotional responses to archetypalmovie clips.
Next, we proceeded to build and evaluate pre-dictive models that quantify these relationships.The evaluation was performed through compari-son of classification accuracies achieved by thepredictive models obtained with linear discrimi-nant analysis (LDA). By feeding a set of recordeddata, LDA would generate a predictive model thatallows us to justify how well these categories ofstimuli can be differentiated according to theemotional responses. In Tables 3 and 4, we demon-strate the confusion matrices of the predictivemodels generated by LDA based on self-reportdata and physiological data. Each row in thesetables shows how well the model can correctly
predict the membership of the movie clips ineach category. The shaded cells indicate the per-centage where the category was correctly pre-dicted, and the emboldened numbers indicate thehighest classification rate among all the predictedgroup members: that is, only those cells withshades and bold numbers are successfully differ-entiated by the statistical model.
We started with the self-report data. The pre-dictive model derived from LDA on self-reportdata about archetypal symbols only obtains24.2% accuracy (cross-validated) and the effectsize is medium (canonical correlation ¼ 0.539).The confusion matrices are demonstrated inTable 3. Among all the seven categories, three cat-egories were successfully differentiated based onthe self-report data: hero’s trials; hero’s return;and mentor. Next, the same analyses on physio-logical data were performed. We extracted usefulfeatures from the raw data of electrocardiographyand skin conductance: heart rate; skin conductancelevel; and skin conductance response. These datawere then fed to LDA for further evaluation. Thepredictive model generated by LDA on the physio-logical features reached an accuracy of 34.4%(cross-validated) and the effect size of the predic-
Archetype
Predicted group membership
Anima Departure Trials Rebirth Return Mentor Shadow Total
Anima 5 6 3 0 21 19 12 66(7.6%) (9.1%) (4.5%) (0%) (31.8%) (28.8%) (18.2%) (100%)
Hero departure 7 16 21 0 5 8 9 66(10.6%) (24.2%) (31.8%) (0%) (7.6%) (12.1%) (13.6%) (100%)
Hero trials 6 14 26 0 6 5 9 66(9.1%) (21.2%) (39.4%) (0%) (9.1%) (7.6%) (13.6%) (100%)
Hero rebirth 10 5 5 0 19 15 12 66(15.2%) (7.6%) (7.6%) (0%) (28.8%) (22.7%) (18.2%) (100%)
Hero return 6 1 4 0 27 16 12 66(9.1%) (1.5%) (6.1%) (0%) (40.9%) (24.2%) (18.2%) (100%)
Mentor 8 2 5 0 11 30 10 66(12.1%) (3.0%) (7.6%) (0%) (16.7%) (45.5%) (15.2%) (100%)
Shadow 8 12 18 0 9 11 8 66(12.1%) (18.2%) (27.3%) (0%) (13.6%) (16.7%) (12.1%) (100%)
Table 3. The confusion matrix of the model obtained from LDA on the self-report data for the movie clips [count (percentage)].Notes: Canonical correlation ¼ 0.539; effect size ¼ medium; 24.2% of the cross-validated grouped cases are correctly classified.Each row in these tables shows how well the model can correctly predict the membership of the movie clips in each category. Thenumbers in bold indicate the highest classification rate for each archetypal image. The shaded cells are correct predictions for eacharchetypal image. Only three categories (Trials, Return, and Mentor) are successfully differentiated by the statistical model.
Grounding the unconscious mind in emotion-driven media
165
Dig
italC
rea
tivity,V
ol.
26
,N
o.
2
Dow
nloa
ded
by [
Ein
dhov
en T
echn
ical
Uni
vers
ity]
at 1
4:58
05
July
201
5
tive model for archetypal symbols is medium(canonical correlation ¼ 0.414). Based on thereview provided in (Novak, Mihelj, and Munih2012), the predictive power of our model obtainedfrom physiological data is in line with other relatedstudies in terms of classification accuracy. Theconfusion matrices are demonstrated in Table 4.It should be noted that all the seven categorieswere successfully differentiated based on physio-logical data, which means that the emotionalresponses measured by physiological signalstoward each category show a unique pattern thatallows a media system to couple the archetypalmovie clips and emotional responses.
The overall classification rates of the self-report data (24.2%) and physiological data(34.4%) toward archetypal movie clips are bothhigher than a chance level (14.29% for seven cat-egories). Moreover, the classification rates of theself-report data are mostly contributed by thethree categories that can be successfully differen-tiated (hero’s trials, hero’s return, and mentor),which are even less than half the number of allcategories. Meanwhile, all the categories can bedifferentiated by the predictive model of physio-logical data at a recognition rate no less than
25%, ranging from 27.3% to 50.0%. It appearsthat the performance of the predictive modelobtained from the physiological data is higherthan the one gained from the self-report data.These results revealed some interesting findingsthat are new to the state of the art in psychology.
6 Discussion
The results of our experiment suggested thatarchetypal movie clips in different categories canbe correctly classified with recognition rateshigher than chance level. This has confirmed theanswer to our first research question. In order toanswer the second research question, it is necess-ary to compare the robustness of the predictivemodels obtained from self-reports and physiologi-cal signals for further inferences about whetherany unconscious emotions were induced by arche-typal media content. Previous studies collectedaffective stimuli based on the categorisation ofbasic emotions, and they reported higher recog-nition rates from self-report data than the resultsof the physiological data (Desmet, Monk,and Overbeeke 2004). While our results of physio-logical data showed a comparable accuracy of
Archetype
Predicted group membership
Anima Departure Trials Rebirth Return Mentor Shadow Total
Anima 33 6 8 5 5 7 2 66(50.0%) (9.1%) (12.1%) (7.6%) (7.6%) (10.6%) (3.0%) (100%)
Hero departure 8 18 8 6 8 10 8 66(12.1%) (27.3%) (12.1%) (9.1%) (12.1%) (15.2%) (12.1%) (100%)
Hero trials 3 9 26 6 8 4 10 66(4.5%) (13.6%) (39.4%) (9.1%) (12.1%) (6.1%) (15.2%) (100%)
Hero rebirth 8 9 3 24 9 6 7 66(12.1%) (13.6%) (4.5%) (36.4%) (13.6%) (9.1%) (10.6%) (100%)
Hero return 5 5 9 8 23 12 4 66(7.6%) (7.6%) (13.6%) (12.1%) (34.8%) (18.2%) (6.1%) (100%)
Mentor 13 7 6 8 7 18 7 66(19.7%) (10.6%) (9.1%) (12.1%) (10.6%) (27.3%) (10.6%) (100%)
Shadow 8 9 16 10 3 3 17 66(12.1%) (13.6%) (24.2%) (15.2%) (4.5%) (4.5%) (25.8%) (100%)
Table 4. The confusion matrix of the model obtained from LDA on the physiological data for the movie clips [count (percentage)].Notes: Canonical correlation ¼ 0.414; effect size ¼ medium; 34.4% of the cross-validated grouped cases are correctly classified.Each row in these tables shows how well the model can correctly predict the membership of the movie clips in each category. Thenumbers in bold indicate the highest classification rate for each archetypal image. The shaded cells are correct predictions for eacharchetypal image. All the categories are successfully differentiated by the statistical model.
Chang et al.
166
Dig
italC
rea
tivity
,V
ol.
26
,N
o.
2
Dow
nloa
ded
by [
Ein
dhov
en T
echn
ical
Uni
vers
ity]
at 1
4:58
05
July
201
5
recognition, the performance of the predictivemodel obtained from self-report data is signifi-cantly lower than previous studies on basicemotions, which is a new finding that relatedstudies have not yet discovered. However, thereare many ways to explain this finding, especiallywhen this might relate to some very fundamentalquestions that have been under debate for years.We consider our results as a preliminary findingso that we have to keep it open for all possibleexplanations.
6.1 Unconscious emotionFrom a methodological standpoint, if we assumethat it is possible to consciously feel emotiontowards archetypal movie clips, a reasonableexplanation would be that the existing dimen-sions of SAM are not comprehensive enough tocover the emotions elicited by archetypalsymbols. The development of the SAM scalemostly focused on quantifying emotions intothree dimensions: valence; arousal; and domi-nance. It appears that these three dimensionsare inadequate for representing emotionsinduced by archetypal movie clips. More dimen-sions are needed for higher-order emotional qual-ities. Some topics related to emotion would beinspiring for new dimensions, such as mindful-ness (Hamilton 2006) or spirituality (Moberg2010). However, the real challenge is to findproper means of measuring these qualities inpsychological studies. Another possible expla-nation is that the self-report technique is not suit-able for measuring emotions other than basicemotions. It might be problematic to ask partici-pants to report their emotions through rationalintrospection because people do not always saywhat they really think (Nisbett and Wilson1977). Moreover, it is suggested that emotionbelongs to the cognitive system of intuition, notthe reasoning part (Kahneman 2003). In thissense, the self-report data collected throughrational reasoning could be a translated versionfrom their intuition, so that these data might bebiased and cannot authentically reflect the trueemotional qualities. Other self-report techniquesthat rely on association might fit better for our
purpose. For example, in the field of psychoana-lysis, guided affective imagery (GAI) is a self-report approach that is designed to observe theparticipant’s unconscious emotional state in away similar to hypnosis (Leuner 1969). Sincethe cinema experience is similar to the processof hypnosis from a Jungian perspective, itmight be a promising direction to apply GAI inthe context of psychological experiments.
Although the above explanations about self-report techniques still require more studies forfurther justification, the analysis on physiologicaldata has shown promising results for measuringemotional responses toward archetypal movieclips. Although the archetypal movie clips in thesame category are edited from different movies,the predictive model obtained from physiologicaldata still allows us to differentiate each category.To some extent these results have provided pre-liminary evidence for the assumption that peoplehave similar bodily responses while beingexposed to archetypal content. Archetypes areunconscious knowledge that cannot be accessedby the conscious mind. However, it can still bemanifested through some unconscious reactions,especially unconscious emotions. In the context ofthe cinema experience, the results of the self-report data collected by introspective reasoninghave indicated the inaccessibility of the consciousmind, but the results of physiological measuresshowed a relatively effective power to couple thearchetypal content with emotional responses. Thisexplanation is also partly in accordance with Dama-sio’s proposition about the distinction betweenemotion and feeling (Damasio 2010). He arguesthat emotions are the phenomena that are carriedout by cognitive actions in our bodies, which canbe observed through physiological signals,whereas feelings of emotions are composite percep-tions of these actions in body and mind. In otherwords, feelings are images of emotions ratherthan emotions themselves. Feelings and emotionsare not always identical, and not all emotions canbe perceived as feelings. Although Damasio’sthesis has been broadly accepted and seems validfor supporting the hypothesis that archetypalmovie clips induce unconscious emotion, it is still
Grounding the unconscious mind in emotion-driven media
167
Dig
italC
rea
tivity,V
ol.
26
,N
o.
2
Dow
nloa
ded
by [
Ein
dhov
en T
echn
ical
Uni
vers
ity]
at 1
4:58
05
July
201
5
an open question for future studies to explorehuman unconsciousness.
The present study did not aim to justify the val-idity of Jung’s theory specifically on its universal-ity. Although our participants had diverse culturalbackgrounds including eleven nationalities, wedid not recruit equal numbers of participantsfrom different countries and most of the partici-pants were from Spain, where the study was con-ducted. Nevertheless, we conducted furtherstatistical analyses specifically on the culturalbackgrounds of the participants in order toanswer this question. There were no significanteffects on the interaction between the categoriesof movie clips and the countries where the partici-pants were born. Although the results to someextent support Jung’s theory, more studies arerequired specifically for assessing the universalityof archetypes.
The validity of emotion recognition acrosspeople has been a difficult problem even in psy-chology, because people may have different reac-tions while experiencing the same emotion(Russell 1994). In our study, we intended toexamine whether the participants have a similarpattern of emotional responses to the given cat-egories of archetypal content. Although theresults suggested that seven categories of archety-pal images can be recognised via measuring phys-iological signals, this finding cannot begeneralised as evidence proving the existence ofthe collective unconscious. Nevertheless, archety-pal symbolism provides a new perspective on ana-lysing media content, which may be valuable fordesigning emotion-driven media systems.
6.2 Implications for media designThe results of our experiment revealed the draw-back of using conscious judgment about emotionalexperience for analysing movies. Previous workon developing media content for emotion-drivenmedia usually relied on the author’s subjectivejudgment on what emotions were induced in thegiven scenes. However, the generated outcomemight be inconsistent because the consciousemotional feeling of the author and the spectatormay differ. Bearing in mind that emotions may
be unconscious, physiological measurementswould be a more reliable approach for validatingemotional responses elicited by the given scenes.In order to confirm the validity of the correlationbetween the scenes and the correspondingemotional responses, it is necessary to conductan experiment as we have demonstrated in thepresent study.
More importantly, the results of our exper-iment have suggested that movie clips belongingto the same categories of archetypal imageinduced similar emotional responses, even ifthese movie clips were extracted from differentmovies. As we pointed out earlier, this theme-based approach focuses on common affectiveelements across different stories, while the story-based approach can only be applied to a specificstory. The theme-based approach allows designersto expand their collections of media content in amore flexible way and the generated outcomewould have richer content, encompassing variousscenes and characters appearing in differentmovies. This approach enables a ‘mash-up’ stylein generating media content for emotion-drivenmedia systems because the archetypal movieclips extracted from different movies deliversimilar symbolic meaning and meanwhile inducemutual emotional responses.
For designing the content of emotion-drivenmedia, archetypal symbolism can be consideredas a design pattern of narratives that are universalto stories in different cultures. Due to the fact thatmany myths, legends and fairy tales in differentcultures share the same structure and components,many scholars in different fields have endeavouredto generalise a universal pattern that fits all kindsof stories (e.g. Campbell 1973; Propp 1984;Levi-Strauss 1955). These structures can still beseen in modern screenwriting (Vogler 2007;Field 1984). Although it is still debatablewhether these patterns cover all stories or moviesexhaustively, these structures can be consideredas a useful reference for designing the content ofmedia systems. For example, the collection ofmovie clips about the ‘mentor’ archetype is auseful resource for designers to create mediacontent that is intended to deliver the feeling of
Chang et al.
168
Dig
italC
rea
tivity
,V
ol.
26
,N
o.
2
Dow
nloa
ded
by [
Ein
dhov
en T
echn
ical
Uni
vers
ity]
at 1
4:58
05
July
201
5
being safe, being guided and kindness. Our studystands as a showcase for utilising archetypal sym-bolism as a design pattern to develop the content ofemotion-driven media systems for the narrativesthat fit the theory of archetypes and the structureof monomyth.
7 Conclusions and future work
These new findings about emotion and archetypalcontent might shed some light on several newdirections both in the design of emotion-drivenmedia systems and research on the unconsciousmind in cognitive science. In terms of design, weproposed a new approach—the theme-basedapproach—for developing media content foremotion-driven media systems. Our studyincluded seven essential archetypes in thepresent study as a starting point and has confirmedthat archetypal content extracted from differentmovies can induce a similar pattern of emotionalresponses across individuals, which was robustenough for emotion recognition. These new find-ings have expanded the possibility of developingnew types of emotion-driven media systems inaddition to the traditional story-based approach.While the results of this preliminary study arepromising, there are more archetypes that are men-tioned in Jungian movie analysis, such as tricksterand shape-shifter. To apply archetypal symbolismin practice, it is suggested that more archetypesshould be included to establish a basis for exper-imental evidence. Although our study has provedthe validity of physiological measurement inrecognising emotional responses to archetypalsymbols, the classification rate is relatively lowfor practical use in design. More movie clipsshould be included in each category in order toenhance the statistical power of the predictivemodel and achieve a higher classification rate.
From a psychological perspective, our studyrevealed new evidence about unconsciousemotion. The research on the unconscious mindhas drawn more and more attention. As atheory that has been used to explain humanunconsciousness, Jung’s model of the humanmind might shed some light on exploring new
directions in cognitive science. We conductedan empirical study according to the psychologicalparadigm and the results showed that archetypalmovie clips in the same category of archetypewere correlated to a unique physiologicalpattern that can be distinguished by computingsystems. While the results are promising, moreevidence is needed to justify the hypothesesthat Jungian scholars have put forward. Fromthe standpoint of cognitive science, our studyprovided more questions than answers. Researchon media content and the unconscious mind ofhumans is worthwhile to proceed with this direc-tion for future studies.
Acknowledgements
This work was supported in part by the ErasmusMundus Joint Doctorate in Interactive and Cogni-tive Environments, which is funded by theEACEA Agency of the European Commissionunder EMJD ICE FPA n 2010-0012.
Notes1 Monomyth is a metaphorical structure representing
mythical narratives about heroes. However, whenthis concept is used in movie analyses, the hero hereis different from the ‘superhero’ that is known as aspecial theme, but can be any characters that gothrough a self-evolved process to become better orbring greatness to others.
2 The one-minute duration was chosen for practicalreasons. First, if the whole experiment takes toolong, the participants might get tired and give biasedresults. Second, the length of the stimuli influencesthe adapting speed of the system. If the length isfive minutes, the media system also needs fiveminutes to start to generate correct predictions whenadapting to a new spectator. Third, although thelength should be as short as possible, it still cannotbe too short to enact emotional experiences.Therefore, one minute appears to be a reasonablechoice.
3 Since the early 1930s, The Archive for Research inArchetypal Symbolism (ARAS), founded byJungian scholar Olga Frobe-Kapteyn, has beencollecting and annotating mythological, ritualisticand symbolic images from all over the world andfrom all epochs of human history.
Grounding the unconscious mind in emotion-driven media
169
Dig
italC
rea
tivity,V
ol.
26
,N
o.
2
Dow
nloa
ded
by [
Ein
dhov
en T
echn
ical
Uni
vers
ity]
at 1
4:58
05
July
201
5
ReferencesAllers, Roger, and Rob Minkoff. 1994. The Lion King.
Film. USA: Buena Vista Pictures.
Bargh, J. A., and E. Morsella. 2008. “The UnconsciousMind.” Perspectives on Psychological Science 3 (1):73–79.
Barrett, Lisa Feldman. 2011. “Constructing Emotion.”Psychological Topics 20 (3): 359–380.
Barrett, Lisa Feldman, Batja Mesquita, K. N. Ochsner, andJames J. Gross. 2007. “The Experience of Emotion.”Annual Review of Psychology 58: 373–403. doi:10.1146/annurev.psych.58.110405.085709.
Bradley, M. M., and P. J. Lang. 1994. “MeasuringEmotion: The Self-Assessment Manikin and theSemantic Differential.” Journal of BehaviorTherapy and Experimental Psychiatry 25 (1): 49–59.
Bradshaw, Sally, and Lance Storm. 2013. “Archetypes,Symbols and the Apprehension of Meaning.” Inter-national Journal of Jungian Studies 5 (2): 154–176.doi:10.1080/19409052.2012.685662.
Broda, Joseph. 1994. “Jung and the Godfather Movies:Analyzing Film from an Archetypal Perspective.”PhD diss., San Jose State University.
Burns, Adrian, Emer P. Doheny, Barry R. Greene,Timothy Foran, Daniel Leahy, Karol O’Donovan,and Michael J. McGrath. 2010. “SHIMMER: AnExtensible Platform for Physiological SignalCapture.” In Annual International Conference of theIEEE Engineering in Medicine and Biology Society,3759–3762. doi:10.1109/IEMBS.2010.5627535.
Cabot, Adele. 2011. “A Review of ‘The King’sSpeech’.” Psychological Perspectives 54 (2): 239–243.
Caldwell, Marylouise, Paul Henry, and Ariell Alman.2010. “Constructing Audio-Visual Representationsof Consumer Archetypes.” Qualitative MarketResearch 13 (1): 84–96. doi:10.1108/13522751011013990.
Campbell, Joseph. 1973. The Hero with a ThousandFaces. Princeton, NJ: Princeton University Press.
Chang, Huang-Ming, Leonid Ivonin, Marta Diaz,Andreu Catala, Wei Chen, and Matthias Rauterberg.2013a. “Experience the World with ArchetypalSymbols: A New Form of Aesthetics.” In Distribu-ted, Ambient, and Pervasive Interactions (LectureNotes in Computer Science Vol. 8028), edited byN. Streitz and C. Stephanidis, 205–14. Berlin:Springer Berlin Heidelberg. doi:10.1007/978-3-642-39351-8_23.
Chang, Huang-Ming, Leonid Ivonin, Marta Diaz,Andreu Catala, Wei Chen, and Matthias Rauterberg.2013b. “From Mythology to Psychology: Identify-ing Archetypal Symbols in Movies.” TechnoeticArts 11 (2): 99–113. doi:10.1386/tear.11.2.99_1.
Chang, Huang-Ming, Leonid Ivonin, Marta Diaz,Andreu Catala, Wei Chen, and Matthias Rauterberg.2013c. “What Do We Feel about Archetypes: Self-Reports and Physiological Signals.” In Proceedingsof 21st European Signal Processing Conference-EUSIPCO2013 (Special Session: Signal ProcessingModels for Influence and Interaction Analysis inCognitive Systems), 1–5. Marrakech: EURASIPOpen Library.
Custers, Ruud, and Henk Aarts. 2010. “The Uncon-scious Will: How the Pursuit of Goals Operatesoutside of Conscious Awareness.” Science 329(5987): 47–50. doi:10.1126/science.1188595.
Damasio, Antonio R. 2010. Self Comes to Mind: Con-structing the Conscious Brain. New York: PantheonBooks. doi:10.1145/1142405.1142428.
Desmet, Pieter M. A., A. F. Monk, and K. Overbeeke.2004. “Measuring Emotion: Development andApplication of an Instrument to Measure EmotionalResponses to Products.” In Funology: From Usabil-ity to Enjoyment, edited by Mark A. Blythe, AndrewF. Monk, Kees Overbeeke, and Peter C. Wright,111–23. Dordrecht: Kluwer Academic. doi:10.1007/1-4020-2967-5_12.
Di Paolo, Ezequiel, Marieke Rohde, and HannekeDe Jaegher. 2010. “Horizons for the EnactiveMind: Values, Social Interaction, and Play.” In Enac-tion: Towards a New Paradigm for CognitiveScience, edited by J. Stewart, O. Gapenne, and Eze-quiel Di Paolo, 33–88. Cambridge, MA: BradfordBooks.
Dijksterhuis, A. P. 2004. “Think Different: The Meritsof Unconscious Thought in Preference Developmentand Decision Making.” Journal of Personality andSocial Psychology 87 (5): 586–98. doi:10.1037/0022-3514.87.5.586.
Ekman, Paul, and D. Cordaro. 2011. “What is Meant byCalling Emotions Basic.” Emotion Review 3 (4):364–370. doi:10.1177/1754073911410740.
Engel, A. B. 1984. “The Brain and Itself: A Study in theArchetype of Existence.” Journal of Social and Bio-logical Structures 7 (3): 207–217. doi:10.1016/S0140-1750(84)80033-0.
Faber, Michael A., and John D. Mayer. 2009. “Resonanceto Archetypes in Media: There’s Some Accounting
Chang et al.
170
Dig
italC
rea
tivity
,V
ol.
26
,N
o.
2
Dow
nloa
ded
by [
Ein
dhov
en T
echn
ical
Uni
vers
ity]
at 1
4:58
05
July
201
5
for Taste.” Journal of Research in Personality 43 (3):307–322. doi:10.1016/j.jrp.2008.11.003.
Field, Syd. 1984. Screenplay: The Foundations ofScreenwriting. New York: Dell.
Fincher, David. 1999. Fight Club. Film. USA: Twenti-eth Century Fox.
Froese, Tom, and Ezequiel Di Paolo. 2011. “The Enac-tive Approach: Theoretical Sketches from Cell toSociety.” Pragmatics & Cognition 19 (1): 1–36.doi:10.1075/pc.19.1.01fro.
Gibson, Mel. 1995. Braveheart. Film. USA: ParamountPictures.
Gigerenzer, G. 2007. Gut Feelings: The Intelligence ofThe Unconscious. New York: Penguin Group.
Gronning, Torben, Patricia Sohl, and Thomas Singer.2007. “ARAS: Archetypal Symbolism andImages.” Visual Resources 23 (3): 245–267.
Hamilton, N. A. 2006. “Enhancing Health and Emotion:Mindfulness as a Missing Link between CognitiveTherapy and Positive Psychology.” Journal ofCognitive Psychotherapy 20 (2): 123–134.
Hannan, David. 2010. Coral Sea Dreaming: Awaken.Film. Australia: Roadshow Entertainment.
Hauke, Christopher, and Ian Alister. 2001. Jung andFilm: Post-Jungian Takes on the Moving Image.Hove: Brunner-Routledge.
Hautala, Tiina, and Vesa Routamaa. 2008. “Archetypesand Types.” In Psychological Type and Culture-East& West: A Multicultural Research Conference, 1–19.
Heyraud, Joyce King. 1998. “Film Review: SavingPrivate Ryan Directed by Steven Spielberg.”Psychological Perspectives 38 (1): 154–157.
Hooper, Tom. 2010. The King’s Speech. Film. UK:Momentum Pictures.
Jackson, Peter. 2001. The Lord of the Rings: The Fellow-ship of the Ring. Film. New Zealand: New LineCinema.
Jackson, Peter. 2002. The Lord of the Rings: The TwoTowers. Film. New Zealand: New Line Cinema.
Jacobi, Jolande. 1973. The Psychology of C. G. Jung.London: Yale University Press.
Jones, Raya A. 2003. “Mixed Metaphors and NarrativeShifts: Archetypes.” Theory & Psychology 13 (5):651–672. doi:10.1177/09593543030135008.
Jung, C. G. 1959. The Archetypes and the CollectiveUnconscious. Translated by R. F. Hull. Princeton,NJ: Princeton University Press.
Jung, C. G. 1964. Man and his Symbols. Garden City,NY: Doubleday.
Jung, C. G. 1967. The Spirit in Man, Art, and Literature.London: Routledge and Kegan Paul.
Jung, C. G. 1969. Aion: Researches into the Phenomen-ology of the Self. 2nd ed. Princeton, NJ: PrincetonUniversity Press.
Kahneman, Daniel. 2003. “Maps of BoundedRationality: Psychology for Behavioral Economics.”American Economic Review 93 (5): 1449–1475.doi:10.1257/000282803322655392.
Kaipainen, Mauri, Niklas Ravaja, Pia Tikka, RasmusVuori, Roberto Pugliese, Marco Rapino, and TapioTakala. 2011. “Enactive Systems and EnactiveMedia: Embodied Human–Machine Couplingbeyond Interfaces.” Leonardo 44 (5): 433–438.
Knox, Jean M. 2001. “Memories, Fantasies, Archetypes:An Exploration of Some Connections between Cog-nitive Science and Analytical Psychology.” TheJournal of Analytical Psychology 46 (4): 613–635.
Kreibig, Sylvia D. 2010. “Autonomic Nervous SystemActivity in Emotion: A Review.” Biological Psy-chology 84 (3): 394–421. doi:10.1016/j.biopsycho.2010.03.010.
Leuner, H. 1969. “Guided Affective Imagery (GAI): AMethod of Intensive Psychotherapy.” AmericanJournal of Psychotherapy 23 (1): 4–22.
Levi-Strauss, Claude. 1955. “The Structural Study ofMyth.” The Journal of American Folklore 68(270): 428–444.
Lim, Lewis K. S., Frank Acito, and Alexander Rusetski.2006. “Development of Archetypes of InternationalMarketing Strategy.” Journal of InternationalBusiness Studies 37: 499–524.
Lindquist, Kristen A., Erika H. Siegel, Karen S. Quigley,and Lisa Feldman Barrett. 2013. “The Hundred-YearEmotion War: Are Emotions Natural Kinds orPsychological Constructions? Comment on Lench,Flores, and Bench (2011).” Psychological Bulletin139 (1): 255–263. doi:10.1037/a0029038.
Martin, Francois, F. Carminati, and G. G. Carminati.2013. “Quantum Information Theory Applied toUnconscious and Consciousness.” NeuroQuantol-ogy 11 (1): 16–33.
Maturana, Humberto, and Francisco Varela. 1980.Autopoiesis and Cognition: The Realization of theLiving. Dordrecht: D. Reidel.
McLoone, H. E. 2010. “Product Archetype of PersonalComputers as an Expression of the Collective
Grounding the unconscious mind in emotion-driven media
171
Dig
italC
rea
tivity,V
ol.
26
,N
o.
2
Dow
nloa
ded
by [
Ein
dhov
en T
echn
ical
Uni
vers
ity]
at 1
4:58
05
July
201
5
Unconsciousness of People on Their Hero’sJourney.” Proceedings of the Human Factors andErgonomics Society Annual Meeting 54 (20):1771–1775. doi:10.1177/154193121005402006.
McTeigue, James. 2005. V for Vendetta. Film. USA:Warner Bros.
Mendes, Sam. 1999. American Beauty. Film. USA:DreamWorks.
Miller, Neal E. 1992. “Some Examples of Psychophy-siology and the Unconscious.” Biofeedback andSelf-Regulation 17 (1): 3–16.
Moberg, DO. 2010. “Spirituality Research: Measuringthe Immeasurable.” Perspectives on Science andChristian Faith 62 (2): 99–114.
Munteanu, Anca, Iuliana Costea, Ramona Palos, andAdrian Jinaru. 2010. “Entering in the Essences ofPersonality-Studies over Archetypes.” Procedia -Social and Behavioral Sciences 5 (2010): 2272–2276. doi:10.1016/j.sbspro.2010.07.448.
Neumann, Erich. 1970. The Origins and History of Con-sciousness. Princeton, NJ: Princeton University Press.
Nisbett, Richard E., and T. D. Wilson. 1977. “TellingMore than We Can Know: Verbal Reports on MentalProcesses.” Psychological Review 84 (3): 231–259.
Nolan, Christopher. 2008. The Dark Knight. Film. USA:Warner Bros.
Novak, Domen, Matjaz Mihelj, and Marko Munih.2012. “A Survey of Methods for Data Fusion andSystem Adaptation Using Autonomic NervousSystem Responses in Physiological Computing.”Interacting with Computers 24 (3): 154–172.doi:10.1016/j.intcom.2012.04.003.
Nunn, C. M. H. 1998. “Archetypes and Memes: TheirStructure, Relationships and Behaviour.” Journalof Consciousness Studies 5 (3): 344–354.
Percival, R. S. S. 1993. “Is Jung’s Theory of ArchetypesCompatible with Neo-Darwinism and Sociobiol-ogy?.” Social and Evolutionary Systems 16 (4):459–487. doi:10.1016/1061-7361(93)90018-M.
Perlovsky, Leonid. 2007. “Neural Dynamic Logic ofConsciousness: The Knowledge Instinct.” In Neuro-dynamics of Cognition and Consciousness, edited byLeonid I. Perlovsky and Robert Kozma, 73–108.Berlin: Springer Berlin Heidelberg. doi:10.1007/978-3-540-73267-9_5.
Picard, Rosalind W. 2003. “Affective Computing: Chal-lenges.” International Journal of Human-ComputerStudies 59 (1): 55–64. doi:10.1016/S1071-5819(03)00052-1.
Propp, Vladimir. 1984. Theory and History of Folklore.Translated by Ariadna Y. Martin and RichardP. Martin. Minneapolis, MN: University Of Minne-sota Press.
Rauterberg, Matthias. 2010. “Emotions: The Voice ofthe Unconscious.” In Entertainment Computing-ICEC 2010 (Lecture Notes in Computer Science,Vol. 6243), edited by H. S. Yang, R. Malaka, J.Hoshino, and J. H. Han, 205–215. IFIP InternationalFederation for Information Processing, Springer.doi:10.1007/978-3-642-15399-0_19.
Robertson, Robin. 2013. “Inner Voices: The Shadowand Other Inner Personalities.” Psychological Per-spectives 56 (3): 289–309. doi:10.1080/00332925.2013.814510.
Roesler, Christian. 2012. “Are Archetypes TransmittedMore by Culture than Biology? Questions Arisingfrom Conceptualizations of the Archetype.” TheJournal of Analytical Psychology 57 (2): 223–246.doi:10.1111/j.1468-5922.2011.01963.x.
Ronnberg, Ami, and Kathleen Martin. 2010. The Bookof Symbols: Reflections on Archetypal Images.Cologne: Taschen.
Rottenberg, Jonathan, Rebecca D. Ray, and James J.Gross. 2007. “Emotion Elicitation Using Films.” InHandbook of Emotion Elicitation and Assessment,edited by James A. Coan and John J. B. Allen, 9–28. Oxford: Oxford University Press.
Russell, James A. 1994. “Is there Universal Recognitionof Emotion from Facial Expressions? A Review of theCross-Cultural Studies.” Psychological Bulletin 115(1): 102–141. doi:10.1037/0033-2909.115.1.102.
Scheirer, Jocelyn, and Rosalind W. Picard. 1999. “Affec-tive Objects.” In MIT Media Laboratory PerceptualComputing Section Technical Report No. 524.
Soleymani, Mohammad, Jeroen Lichtenauer, ThierryPun, and Maja Pantic. 2012. “A Multi-Modal Affec-tive Database for Affect Recognition and ImplicitTagging.” IEEE Transactions on Affective Comput-ing 3 (1): 42–55. doi:10.1109/T-AFFC.2011.25.
Soleymani, Mohammad, Maja Pantic, and Thierry Pun.2012. “Multi-Modal Emotion Recognition inResponse to Videos.” IEEE Transactions on Affec-tive Computing 3 (2): 211–223. doi:10.1109/T-AFFC.2011.37.
Tikka, Pia. 2004. “(Interactive) Cinema as a Model ofMind.” Digital Creativity 15 (1): 14–17. doi:10.1076/digc.15.1.14.28151.
Tikka, Pia. 2006. “Cinema as Externalization of Con-sciousness.” In Screen Consciousness: Mind,
Chang et al.
172
Dig
italC
rea
tivity
,V
ol.
26
,N
o.
2
Dow
nloa
ded
by [
Ein
dhov
en T
echn
ical
Uni
vers
ity]
at 1
4:58
05
July
201
5
Cinema and World, edited by M. Punt and R. Pepper-ell, 139–162. New York: Rodopi.
Tikka, Pia. 2010. “Enactive Media – Generalising fromEnactive Cinema.” Digital Creativity 21 (4): 205–214. doi:10.1080/14626268.2011.550028.
Tikka, Pia, Rasmus Vuori, and Mauri Kaipainen. 2006.“Narrative Logic of Enactive Cinema: Obsession.”Digital Creativity 17 (4): 205–212. doi:10.1080/14626260601074078.
Tornatore, Giuseppe. 2000. Malena. Film. Italy:Miramax Home Entertainment.
Tsai, Shu-pei. 2006. “Investigating Archetype-IconTransformation in Brand Marketing.” MarketingIntelligence & Planning 24 (6): 648–663. doi:10.1108/02634500610701708.
Tykwer, Tom. 2006. Perfume: The Story of a Murderer.Film. Germany: DreamWorks.
Van den Broek, Egon L. 2013. “Ubiquitous Emotion-Aware Computing.” Personal and UbiquitousComputing 17 (1): 53–67. doi:10.1007/s00779-011-0479-9.
Varela, Francisco J., Evan T. Thompson, and EleanorRosch. 1992. The Embodied Mind: CognitiveScience and Human Experience. Cambridge, MA:MIT Press.
Vogler, Christopher. 2007. The Writer’s Journey: MythicStructure for Writers. 3rd ed. Studio City, CA:Michael Wiese.
Wachowski, Andy, and Lana Wachowski. 1999. TheMatrix. Film. USA: Warner Bros.
Wachowski, Andy, and Lana Wachowski. 2003. TheMatrix Revolutions. Film. USA: Warner Bros.
Walters, Sally. 1994. “Algorithms and Archetypes:Evolutionary Psychology and Carl Jung’s Theoryof the Collective Unconscious.” Journal of Socialand Evolutionary Systems 17 (3): 287–306.
Westen, Drew. 1999. “The Scientific Status of Uncon-scious Processes: Is Freud Really Dead?” Journalof the American Psychoanalytic Association 47 (4):1061–1106.
Wilson, T. D., and Yoav Bar-Anan. 2008. “The UnseenMind.” Science 321 (5892): 1046–1047. doi:10.1126/science.1163029.
Woodside, Arch G., Carol M. Megehee, and SureshSood. 2012. “Conversations With(in) the CollectiveUnconscious by Consumers, Brands, and RelevantOthers.” Journal of Business Research 65 (5):594–602. doi:10.1016/j.jbusres.2011.02.016.
Zehnder, Sean M., and Sandra L. Calvert. 2004.“Between the Hero and the Shadow: DevelopmentalDifferences in Adolescents’ Perceptions and Under-standing of Mythic Themes in Film.” Journal ofCommunication Inquiry 28 (2): 122–37. doi:10.1177/0196859903261797.
Huang-Ming Chang is a PhD candidate affiliatedwith Eindhoven University of Technology and theTechnical University of Catalonia. His researchinterests include human emotion, interactiondesign and digital media.
Leonid Ivonin is a PhD candidate affiliated withEindhoven University of Technology and theTechnical University of Catalonia.
Marta Dıaz is a psychologist, full professor at theDepartment of Management and Director of theInteractive Behaviour and User Experience Labat the Technical University of Catalonia. Herresearch interest is human behaviour interactingwith smart systems.
Andreu Catala is a full professor of the TechnicalUniversity of Catalonia in the area of systemsengineering and automatic control and Directorof the Technical Research Centre for DependencyCare and Autonomous Living. Research interestsare machine learning and soft computing,ambient intelligence and assistive technologies.
Wei Chen is an assistant professor at the Depart-ment of Industrial Design, Eindhoven Universityof Technology. Her research interests includesensor systems for ambient intelligent design,healthcare system design using wearable sensors,wireless body area networks and smart environ-ments.
Matthias Rauterberg is full professor for inter-active systems design at the Department ofIndustrial Design, Eindhoven University ofTechnology. He has over 400 publications inthe areas of human computer interaction, enter-tainment and cultural computing, as well asinteraction design.
Grounding the unconscious mind in emotion-driven media
173
Dig
italC
rea
tivity,V
ol.
26
,N
o.
2
Dow
nloa
ded
by [
Ein
dhov
en T
echn
ical
Uni
vers
ity]
at 1
4:58
05
July
201
5