Commented City Walks


By Jean-Paul Thibaud

This paper presents an interdisciplinary method designed to study urban ambiances. The main goal of the commented city walks approach is to gain access to the in situ sensory experience of passers-by. The key is to acquire accounts of perception in motion. For this, walking, perceiving and describing are simultaneously required. Commented walks are based on three central hypotheses: the situationally-rooted nature of perception, mobility as a condition for the existence of perception, the interlacing of words and perception. This method is applied here to the Grand Louvre in Paris.

Research into description as a privileged means of access to social reality has flourished over the past thirty years (Ackermann W. & al. 1985 ; Quéré L. 1992 ; Revue Enquête 1998). Numerous different qualitative approaches in the humanities have attempted to analyse social life from a sensorial perspective and to situate these using detailed accounts. In the case of human ethnology, ethnography of communication, sociology of communication, ethno-methodology or conversational analysis, this primarily involves observing and writing. These approaches reject the dualism of being and appearing to be in favour of studying social links from what can be seen, heard and reported. The problem is one of putting what appears into words and commenting the world as it unfolds. What is the role of perception in the social construction of reality? What is the extent of the sensorial nature of the social realm? These are the questions that run through the most complex reflections on this subject.

The theme of public space crystallizes certain theoretical, methodological and epistemological considerations in descriptive approaches. As a conceptual category, the “public” links back to the phenomenal nature of reality and the facility of appearing to be through the eyes of others and of perceiving together (Arendt 1961). What becomes public comprises what may be perceived by everybody via his/her senses, what may be instituted by a “common appearing” (Quéré 1991). As such, social life may not be reduced to a sum of singular experiences; it mobilises a community of perceptions and develops on the basis of behaviour that may be observed by everyone. This inter-subjective dimension of public space is based on the fact that others take account of my presence, manifest their actions and enjoy the same access as me. Moreover, perception only makes sense in the context of the places and circumstances in which it unfolds. The use we make of our eyes and ears depends on the micro-ecological framework of the encounters and events with which we are confronted. The possibilities for access to others involve both physical conditions of visibility and audibility and perceptive structuring of sensorial data. Put another way, attention distribution, exposure to others and organisation of perspectives take place based on what it is possible for us to see and hear.

Most sociological approaches that deploy a descriptive approach reserve a central role for context-based social behaviour. Intricate accounts of ordinary gestures, postures, looks and words are usually an essential stage of field research. In particular, audio or video recordings make it possible to conserve a material trace of the activity in situ, to consult the documents as often as one wishes, to analyse several documents together and harness the possibilities offered by the technical medium. Viewing a piece of video without the sound, in slow motion or in freeze-frame mode, or using close-ups helps in analysing phenomena that could only be seen with great difficulty by the naked eye. As regards sound, systematic, repeated listening to recorded conversations makes it possible to analyse the manner in which verbal interactions are structured, either in terms of the prosody of the articulated language (vocality) or how series of words come together (sequentiality). One of the key contributions of these observations/descriptions is to highlight the contextual nature of social behaviour. These approaches constantly illustrate how behaviour only makes sense when placed in the context in which it appears, i.e., the spatial-temporal circumstances, actions in progress, and the expressions and behaviour of others.

Paradoxically, very little research has attempted to gauge the effectiveness of the sensorial environment in terms of how practical actions take place. Some studies offer detailed illustration of how the sensorial environment interferes with and helps to structure the reciprocal perception of urbanites [1] . Indeed, the focus on behavioural adjustments and cooperation and social coordination processes favours an analysis of actors’ interactional skills. This pragmatic approach questions how urbanites seek to make their behaviour clear, observable and intelligible, however it frequently neglects the sensorial context of this reciprocal accessibility (brightness and noise in the place in question). Everything happens as if there was maximum permeability of the sensorial milieu and the perceptive constraints and potentialities of the place were negligible. In other words, the hypothesis of the endogenous organisation of practical action tends to mask the physical, material and environmental conditions of social interaction. Erving Goffman undoubtedly has a special place in this type of approach. His constant focus on the micro-ecological frameworks of interactions led him to develop a conceptual apparatus that specifies the impact of place on actors’ perceptive behaviour (notions of social situation, region, framework, etc.). His monumental work sketches out the possibility of differentiating between places in terms of conditions of access to others. However, the characteristics of the place are usually analysed in terms of physical obstacles – all manner of walls and partitions – that function as barriers to perception, giving rise to a series of oppositions (stage/backstage, opening/closing, manifestation/dissimulation) that provide little indication of the complexity of the environs. Only rarely are we provided with a detailed illustration of how the built environment generates phenomena of reduced, contrasted or bloated observability. Furthermore, instead of focusing on the switch from backstage to centre stage, we wish to understand how the sensorial milieu helps to “showcase” life acted out in public, and how the sensorial qualities of a place are bound up with its public character. So how can we analyse the perceptions of passers-by in public? What is observable data in terms of ambiance?

In attempting to address these questions, I have initiated a methodology for commented city walks adopted and developed by the CRESSON research laboratory (centre for research on sonic space and urban environment). The main goal of the commented city walks approach is to gain access to the sensory experience of passers-by. The key is to acquire accounts of perception in motion. For this, three activities are simultaneously required: walking, perceiving and describing. As we will show below, this method is part of a wider interdisciplinary approach which draws on the engineering sciences (acoustics, lighting technology), design science (architecture, urbanism) and the social sciences (micro-sociology). In situ descriptions are nonetheless central to this approach: it is from such descriptions that hypotheses about ambience phenomena are formulated; further, they are the preferred means for interpreting various corpora. Commented walks are based on three central hypotheses.

1. Perception in context

Contemporary epistemology constantly claims that it is impossible for researchers to achieve a global perspective of their research topics. Regardless of the nature of an observation, we need to contextualise the conditions under which it was made possible. We first need to analyse this contextualisation in specific terms. The notion of context is twofold: both in terms of the methodological clause (the interaction between observable data and observation conditions) and of the research topic (ambience as a sensory context). What consequences can we draw from this dual contextual situation? For one, sensory ecology (Gibson 1986) reminds us that a subject’s perceptive orientations cannot be dissociated from the “affordances” of an environment. As such, perception is not so much based on an environment as it is subject to an environment; it thus needs to be connected to the properties and characteristics of the site studied. Once we admit that environmental context has an impact on perception, the latter can no longer be studied in an abstract manner but rather needs to be examined as part of a perception-environment pairing. Moreover, examining the sensory activity of city-dwellers requires that we get as close as possible to their point of view. They use their senses based on and according to a pragmatic context: taking one’s place in a queue, avoiding collisions on a journey, crossing an intersection on foot, waiting for an acquaintance in the street or just trying to find one’s way are all practices which affect how we perceive things in public (Thibaud 2002). From this perspective, the way things are perceived cannot be dissociated from the course of action in which passers-by are engaged. In general, sociologists offer a second perspective in which they assume the role of a disinterested observer, external to the situation. And yet, given that city-dwellers themselves are observers of public life, can we not instead draw on their ability to observe and describe things? We wish to move away from academic and distanced observation and towards ordinary and engaged description. As such, the description of what is perceptible would no longer be dictated by the researcher but directly by the passer-by. Using the reflexive resources of passers-by provides an opportunity to contextualise their descriptions. Taken as such, the empirical study of urban ambiences requires us to examine the sensory environment of places, the perceptive behaviour of passers-by and the social activities in which they are involved. Ultimately, the aim is to point up the situationally-rooted nature of perception in order to develop an in situ approach.

2. The inevitable “movement” of perception

Whether we refer to phenomenology, sensory ecology or neurophysiology, it now seems illusory to want to dissociate perception and movement. All perception involves “movement”, regardless of how small, to make the act of perception possible. As such, phenomenology is constantly pointing up the founding unity between “feeling” and “moving” (Straus 1935), the chiasmus forged by perception and movement (Merleau-Ponty 1968). Rather than starting with the duality between object and subject, action and perception, we can look at their co-affiliation and concede that there appears to be something akin to an “appearance driver”. From this perspective, movement cannot be reduced to simply changing place or moving from one spot to another; it calls up and reveals sensory qualities: “this amounts to saying that phenomenalisation originates in the world in which the subject is involved through his or her movements, or that it is the moving subject who, by going towards the world, makes it appear” (Barbaras 1992 p.35). Setting the body in motion involves both a practical occupation of the world and an awareness of it. The principle of moving perception is not only based on an ontology of the flesh or a praxis of the perceiving body; it is also useful for understanding the sensory construction of public space. “Movement” is also a contingent possibility for the public sphere. Indeed, it is based on the plurality of perspectives and thus requires a variety of positions and points of view. Paradoxically, public space only begins to exist as a shared world, with unity and shared identity, once it has been seen from a variety of angles. In other words, for public space to exist, I must be able to put myself in another person’s place and to change perspective. Finally, urban ambiences raise the issue of movement on two levels: in terms of motivity, as a condition for the existence of perception, and in terms of mobility, as a condition for the existence of public space. As such, rather than take a static approach, we wish to place movement at the very heart of our field research. To understand the city-dweller’s sensory experience, we shall use a protocol for moving about in the city. Our research is based not only on situated perception but also on perception in motion.

3. The interlacing of words and perception

Western philosophy has a long tradition of opposing what is sensory and what is intelligible: perception on the one hand and concept on the other. Although the two indeed seem very different, we can nonetheless suppose the existence of close – albeit complex and variable – ties between the two registers. We believe that it is possible to assess perception based on what is verbally reported. We cannot of course perceive everything that our senses encounter, but only that which we can put into some sort of form. This is the distinction between sensing and perceiving. On a basic level, our sensory experience is based on a collection of unstable and undefined stimulations of which we have no memory; this is what Leibniz calls “the wave of small, indiscernible and vague perceptions” (1966, p.38). Apperception begins only when we let a gestalt emerge and we release figures from the sensory world. An individual’s relationship with the environment is not purely reactive or reflexive; it involves a degree of configuration by the perceiving subject. Thus, spoken language is not only an instrument which allows a lived experience to be recounted after the fact, represented and shared with others; it actually participates in the experience fully and immediately. Hermeneutics as such insists on the essential linguistic condition of all human experience. [2] Taking a completely different approach, conversation analysis points up how concepts inform perception and how “classes and categories allow us to see” (Sacks 1992). In other words, all experience is conceptually built and is deciphered only through and in the language that it solicits. Events and phenomena perceived in situ are individualised on the basis of conceptual schema which give them shape and meaning. The descriptions we collected during our fieldwork give us access not only to the experience of passers-by but also reveal the configuration operations bound up with the ambience of a place. Our initial assumption is that a sensory environment can work as “speech trigger” and that local ambiences can be grounds for verbalisation. Just as words only take on meaning alongside a non-verbal backdrop, what is perceived inevitably involves a backdrop of inattentiveness. Describing an event presupposes that it was perceptible to us and that it became poignant enough to talk to us, to make us talk and to talk through us.

The entire approach is represented in the diagram below. We will then go on to describe the different phases individually.


Descriptions of perception in movement

Following a preliminary site observation phase (i.e. becoming familiar with the pedestrian network, the site’s architecture, informal observation of social behaviour), we begin collecting reports of perception in movement. Passers-by (regular and non-regular site users) are asked to participate in the in situ experience. The ordinary descriptions provided by passers-by are what constitute the basic corpus for our analysis. This experience involves taking a walk and describing what is perceived and felt during the excursion.

Conditions of the experience

The research protocol is based on three types of instructions which structure the experience:

  • Instructions related to the description: describe, as precisely as possible, the ambience of the place as it is perceived and felt during the walk. All sensory modalities can be used: visual, acoustic, tactile, heat, olfactory, kinaesthetic, etc. In order to know where an individual is during their description, they are asked to regularly provide spatial references, useful during the analysis phase.
  • Instructions related to the route: he investigation site is established ahead of time, but the actual route taken can be left to a certain extent up to the interviewee. If they wish, the passer-by can stop momentarily, go back to a place or change pace.
  • Instructions related to the conditions of the experience: given the focus and attention required for such an experience, the walk lasts roughly twenty minutes, but may be extended if the person so wishes. All comments are fully recorded using a portable recording device. The route is travelled with the researcher who records the description but intervenes as little as possible and just tries to be a friendly listener, possibly restarting the conversation if the walker-observer is clearly having trouble.

During a second stage, the trajectory is reconstructed on paper with the participant. A map is useful when the places visited are particularly complex and it is difficult to reconstitute the route. If this is not the case, it is preferable to conduct the reconstitution without a graphic aid in order to encourage the use of memory (What are places called? What do we use to recall them? etc.). Then a short interview is conducted in which the following issues are tackled: possibility of distinguishing between the places visited according to their ambiences, dividing the walk into sequences, the most striking events during the walk, familiarity with the neighbourhood and how often it is visited, assessment of the experience, personal information (age, profession, site use frequency, etc.).

The experience is repeated roughly twenty times with different people who are contacted either through personal networks or directly at the fieldwork site (in this case, passers-by are engaged in a specific course of action: they are going to work, to take public transportation, for a walk, shopping, etc.). In order to ensure that the population and activity contexts are as diverse as possible, it is important not to rely solely on personal contacts.

If there is spatial coherence in the corpus (the same site for all of the commented walks), the search for diversity in the descriptions is based on three variables:

  • A variety of routes. If left up to the walker-observer, the route chosen will not necessarily be the same for each person. Such spatial variations are interesting for more than one reason. Aside from the fact that they tend to cover a good part of the research field, they provide precious clues about how urban space is appropriated. They also allow for a comparison of descriptions according to the walker’s orientation (directionality of the chosen route) and how a place was accessed (environs of squares and intersections).
  • A variety of circumstances. The sensory context of a place changes over the course of a day or week (as well as with the seasons). Descriptions of the same walk differ based on whether it was taken during the day or at night, whether it was a sunny day, taken alone or with other members of the public, during organised events or during quiet periods. Therefore, it is best to diversify the temporal conditions of the experience as much as possible. These two aspects – directionality of the route and temporality at the site – are the two most important aspects of the description context to consider during the analysis phase.
  • A variety of points of view. This will depend on the people participating in the experience. Aside from the traditional variables such as age and sex, there are three other parameters at play when selecting walker-observers: socio-cultural status which produces specific verbalisation processes, how well a site is known which will engage memories to variable degrees, and a person’s visitor status (ordinary passer-by, tourist, local shopkeeper, homeless person, etc.) which will often lead to implicit representations. It is important to compare and contrast points of view, to tease out the similarities that go beyond the differences and reinterpret similar content from unique ways of describing them. Such cross-analysis is a means of reconstructing the inter-subjective dimension of the experience and of showing how a site can generate shared perceptions.

Ways of describing, means of perceiving

The descriptions recorded through this research technique are transcribed and aim to remain as faithful as possible to the fluctuations in recorded speech (e.g. pauses, probing, hesitation, onomatopoeia and stammering). The great diversity of comments means that the descriptions must be read repeatedly in order to progressively extract the perceptive “moorings” rooted therein. In this respect, analysing the comments is not so much about classifying the objects perceived during the walk (what is perceived?) as about examining how people relate what they perceive (how do we perceive?). Although they are not devoid of interest, we tried to move beyond two immediate observations. For one: the filter of representations in descriptions. We observed that some comments implicitly or explicitly contain personal judgements which guide people’s comments. The way things are described is influenced by our feelings about (positive or negative), and by our use of a place (a tourist discovering a new place, the banality of everyday routine, a commercial strategy, etc.). Secondly: use of different types of discourse. Descriptions take on a different form depending on whether we are observing, qualifying, detailing, clarifying or listing. And yet beyond such purely descriptive aspects, there are also attempts to explain (hypotheses, reasoning) and assess (aesthetic or functional).

In addition to a sensory interpretation, several descriptors were selected during the analysis. Through different means, each provides access to situated perceptions and sensory contexts regarding the places visited.

a Spatial-sensory associations. Very often, passers-by call on their perceptive memory and describe the ambience of a site they are visiting by association. Some urban spaces thus work as veritable sites of sensory reference. Associating an ambience perceived in situ with places like a “train station”, “airport”, “swimming pool”, “Parisian passageway” or “greenhouse” provides insight into the acoustic nature, lighting and temperature in a place. Such indications become explicit when the observer explains or justifies the relevance of the metaphors used.

b Perceptive transitions. Another means to describe the ambience of a place is to note the obvious changes which occur. For example, we can note that in some places “it is calmer”, “voices sound more distant”, “there is less light”, “it is brighter”, “it is too hot”, etc. By highlighting differences in intensity and changes in quality, the observer describes an ambience in their spatial-temporal setting according to the route taken and specific circumstances. Whether related to an event and/or a location, such perceptive transitions allow us to identify the sensory connections that exist between places.

c The verbal field of appearance. Describing what we perceive not only provides an inventory of facts and obvious details that can be heard and seen. The use of verbs such as “seems to”, “appears to” and “looks like” expresses a degree of perceptive uncertainty and ambiguity. As John Austin (1964) has shown, such verbs are not always interchangeable; they expose the dilemma between a matter of fact and a question of language. By listing the circumstances in which such verbs are used, we can identify situations which are problematic from a perceptive point of view. The analysis must not be limited to understanding when others become visible or audible, but rather specifically how and under what conditions. A certain number of verbs provide invaluable information on contexts of observability and types of access to others: “to contrast”, “detach oneself”, “merge into”, “be outlined against”, “emerge”, “be hidden”, etc. Such terms are particularly interesting given how they describe the way the public is present by connecting the sensory details of space to the observer’s configuring activity. More specifically, they allow us to name perceptive phenomena, understand how a form emerges from within the whole and identify different types of figure/background relationships.

d Reflexive formulations. The last point used during the analysis involves the observer-walker’s reflexive wording. This helps clarify the two main activities of the experience: on the one hand, the perceiving subject’s perceptive orientation (e.g. “I’m looking up”, “I’m turning around”, “I’m listening”); and, on the other hand, the walker’s momentum (e.g. “I’m hesitating”, “I’m accelerating”, “I’m stopping”, “I’m drawn towards”). Such word choices indicate the possibilities available to passers-by to adapt to the circumstances at a given moment and potentially impact the practical context of the experience. As part of the public, observers have to respond to the situations and events they encounter along their way.

These different types of descriptive behaviour are not independent of each other; rather, they correspond with different ways of expressing an in situ experience. Given that they concern a single site, each rounds out the others and provides a possible breakdown of the practical-sensory field.

“Polyglot compilations”

Thus far, we have focused on the first phase of analysis which involves sorting and selecting a certain amount of descriptive data. In order to be fully operational, the data must undergo several additional operations. Firstly, they need to be replaced in the flow of the description according to what is said before and after. Most often, it is the connections between proposals that allow us to understand the meaning of each. Then we need to check the inter-subjective content of comments by doing cross-checks and comparisons. The recurrence and redundancy of similar comments from different observers is proof of a certain perception community. Finally, the descriptions are organised according to the places visited and the conditions involved. This provides us with a collection of information and hypotheses about the sensory-motor phenomena related to a visited site. The categories and terms employed to describe such phenomena are largely based on research conducted by the CRESSON [3] research unit.

Reconstructing the walk

The challenge then is to synthesise all of the descriptions compiled during the first phase of fieldwork. To do so, we reconstruct the narratives into an “ideal” walk (in the sense of Max Weber’s ideal type). For a given walk, we need to reconstruct the route which brings out the existing spatial-perceptive potentiality most effectively. Therefore, the difficulty is not reconstruction verisimilitude but its demonstrative strength. We do not actually do a complete re-writing, but rather a collage of fragments from the different descriptions. Changes in the person speaking are identified by a full stop. These “polyglot compilations” – a patchwork of diverse visitor comments – nonetheless follow the same trajectory logic. Two basic rules regulate this descriptive compression: fragments which best reveal the local sensory context should be selected, and the speaker’s localisation and the route’s directionality must be respected. The “reconstructed description” (in the column on the left) is accompanied by a “reader’s guide” (in the column on the right) which indicates the sensory-motor phenomena identified. The “polyglot compilation” below can be read like a descriptive analysis or a result in progress. Mid-way between description and analysis, it gives voice to a place whilst shaping understanding.

The example below is drawn from part of a walk at the Grand Louvre in Paris. The walk begins in the courtyard around the pyramid and continues through to the Louvre’s Grande Galerie.

Excerpt from a “polyglot compilation” at the Grand Louvre

I am on a large and very sunny esplanade, you can feel the heat rising from the ground; I am drawn towards this pyramid like an open, central form towards which you are almost obliged to move, I mean, there is a very strong attractive force in the middle of this square. It’s true that it’s attractive, it really catches the eye. Here, you really feel like you’re in the city, with traffic all around, the sound of people – albeit quite faint – talking, entering, waiting in the queue. The noises are quite faint, I can mostly hear children’s voices, a bit like in a public park.

We are going to enter the Pyramid. So we’re now entering, there, the sound has already completely changed since we’ve left all of the cars behind us, the sound is very muffled, it resonates a lot, like when you enter a railway station actually, or an airport, it’s a bit like that kind of place. The noise is what’s most striking here, it’s filled with noise that gives you the impression that you’re in a closed space. Here, there is a kind of muddled noise, the same noise you hear in swimming pools, you can hear everything and nothing at all because you can’t actually understand anything or distinguish things. The sound of people is less precise than outside, there is a sort of background noise, a hum. You really need to strain your voice when you talk here, to make yourself audible; although the sound does not seem all that loud it is actually very present. So now, it’s really like a greenhouse in here, there is a concentration of heat. Under the pyramid, we really feel much more stifled by the heat. Now my eyes automatically look up to the sky, because the sky looks really unique, through this glass roof, which outlines a piece of blue sky, although it was already there when we were outside, but I was not looking at it. So of course thereafter, when we move ahead, we’re drawn towards a shaft. I think that this platform was designed for us to stop. Here, we are overlooking the large arena – it actually looks like a large arena – of the Louvre museum’s entrance, under the pyramid; so there are people. The people are down there, it is a sort of small closed zone, a kind of… I don’t know… micro-society. You usually get the impression of a living painting, an anthill, where you can see people waving to each other, meeting up, queuing for tickets. What I really like are the materials used which reflect, and are bright. The use of white marble.






Warming up





Ambience at the project scale

The text above highlights different sensory phenomena expressed during the walk. Such an intermediary result still needs to be worked on. The polyglot compilation – the intermediary step in the analysis – now needs to undergo a secondary description which clarifies and contextualises the sensory-motor effects detected. We need to not only make the terms in the column on the right understandable, but also re-contextualise the sensory-motor effects previously identified. Indeed, these are not necessarily stable or permanent; they can vary over time according to the number of visitors, the relative amount of natural light (daytime/night time ratios) and events at a given time. It is thus necessary to return to the field. The analysis of the comments obtained during the first phase of fieldwork provides a structured framework for observation of the site. The goal is to precisely identify the conditions under which the phenomena described by the passers-by emerge. The relationship between observation and description is thus reversed: the goal is no longer to describe what is perceived but to link the descriptions to what is observable on site. Aside from ethnographic observations which allow us to describe social behaviour and preserve a physical trace of the phenomena identified (recorded sound and images), we take architectural, acoustic, lighting and thermal and aeraulic measurements which subsequently allow us to analyse the physical nature of the perceived effects. Although we shall not develop this last point here, this is an integral part of the research method. [4]

Contextualising phenomena

The final processing of the different corpora obtained in the second phase of fieldwork allows us to not only clarify the conditions and circumstances under which phenomena appear, but also to analyse the built infrastructure in terms of ambience. The challenge is then to understand how each phenomenon detected fits together and works in combination with the others to produce local sensory configurations. This analysis of a place’s ambience is based on three additional approaches. The first is that of the built infrastructure. The goal is to describe the physical makeup of the place. More specifically, this first category aims to expose the interaction between the built environment and physical signals. The phenomena detected here are measurable. The second approach concerns the sensory features. The goal is to describe the perceptive component of the in situ experience. The phenomena detected here can be expressed. The third approach deals with the social features. The goal is to describe the public dimension of visitors’ behaviour. The phenomena detected here are observable. Let us take the example of the Hall Napoléon in the Grand Louvre (the space located under the pyramid) to illustrate how data can be reinterpreted this way. Only the visual and acoustic aspects will be examined here.

Seeing and hearing in the Hall Napoleon

The basement as a bright background. Visually, the Hall Napoleon can best be described as a place of “overexposure”, meaning that people are extremely susceptible to be seen by others. In other words, visitors situated in the basement are particularly visible to those looking down from the entrance level above. Everything is done to encourage those arriving to look at what is happening below. The layout almost exhorts visitors to observe.


  • Visual attraction. From the outside, the monumental emergence of the pyramid works as a veritable channel of curiosity and attraction. Everything seems to exhort passers-by to enter the large glass space. Once inside the pyramid, the attraction effect persists and draws people’s attention towards the basement: “we are drawn towards a shaft.” Visitors become momentarily captivated by the view, which dominates the foyer below. Not only do they need to take their time to identify how to access the basement, but they need to adapt their trajectory and go either to the right (escalator) or to the left (spiral staircase). The staircase and escalator are off-centre in relation to the entrance and break up the continuity of movement. However, the view is directly in line with the entrance and calls out to visitors, inviting them to pause.
  • A commanding view. Upon entry, visitors are well above the foyer located in the basement (9 meters below); this elevation provides a commanding view of what is happening below: “we are overlooking the large arena”. This dominant perspective allows the visitor to survey the entire scene in the foyer whilst remaining detached from the activity below. Further, the foyer’s closed nature and the mezzanines along almost all sides accentuates the panoramic view and closely resembles an arena.
  • Figure-backdrop contrast. The basement does not contain any furniture or visual obstacles and is thus a backdrop against which only the outlines of passers-by appear: “people clearly stand out from the floor”. From the lookout up at the entrance, there is nothing to mask their presence. The brightness of the walls and the natural light flowing through the glass pyramid accentuate the luminosity in the foyer and make the people in the basement even more visible. Depending on the amount of sunshine, the luminance of the basement varies between 100 cd/m² and 35 000 cd/m². Seen from above, human forms are even more striking as they are reflected off the floor which is entirely light in colour.

The invitation to visit, the commanding perspective and the uniformly light background mean that visitors in the basement are overexposed. This visibility context is nonetheless subject to variation. For one, the number of visitors in the foyer affects visibility. When the area is relatively empty, each person becomes a visible entity that clearly stands out from the others; each is optimally exposed. At busy times, however, the forms which stand out correspond more to clusters of fluctuating groups and queues. Each individual becomes an indistinct element in a collective whole. Further, visibility also varies according to the lighting. When the sun is shining, natural light makes the stone appear brighter and accentuates the contrast between the bright surface in the basement and the visitors. The shaded areas in the basement (and, similarly, the exclusively artificial lighting at night) produce less contrast than the areas directly lit up by the sun. However, too much reflection from the sun on the floor can cause a glare and make it difficult for visitors up above to prolong their observation.

The indistinct hum in the background. Acoustically, the Hall Napoleon is filled with an indistinct hum into which visitors are immediately submerged. When arriving from outside, visitors immediately notice a distinct change in the acoustic ambience. The muddled ambient noise reinforces the sense of a collective presence yet conserves the impersonality of contact with others. Once again, several phenomena work together to produce this acoustic immersion.

  • The reverberation effect. Composed of particularly reflective matter (stone and glass), there are several reasons why this large, closed space (35 000m3) is particularly prone to strong reverberation: “there is a strong resonation here”. Sounds bounce off the surfaces in the surrounding environment and persist even after the source of a sound has stopped. The larger the crowd, the louder it sounds and the greater the indistinct hum which affects how noise is distinguished. Without visitors, the reverberation time in the foyer is close to 6 seconds (TR60).
  • The humming effect. Several mechanical sources (escalators, ventilation) generate continuous noise of equal intensity which accompanies visitors throughout their stay: “a sort of permanent noise”. The sound of such equipment is propagated throughout the space with a slight attenuation based on distance. This humming effect sets the tone of the site’s ambience by creating a constant and uninterrupted acoustic backdrop. There is no pause in sound to let the ear relax and momentarily escape the acoustic continuum.
  • The mask effect. The background noise under the pyramid is relatively loud and varies between 70 dB(A) and 72 dB(A) with visitors. This acoustic reality tends to reduce the intelligibility of speech, mask visitors’ voices and require them to raise their voices: “you need to strain your voice when you talk here”. Only a few higher pitched sounds (children’s screams, ringtones, etc.) are decipherable.
  • The mixing effect. The noise produced by the mechanical equipment (ventilation and escalators) blends with voices from the public to such an extent that visitors have trouble clearly dissociating the two types of noise. Speech also tends to merge into the site’s acoustic backdrop: “voices become a kind of mechanical hum”. Acoustically, the background noise spectrum is dominant between 250 Hz and 1 KHz, a spectrum close to that of speech.

The four effects identified all contribute in one way or another to the acoustic immersion of visitors. This immersion can be more or less pervasive depending on the density of people in a given space. We observed, for example, that the presence of visitors is an important factor in the loudness rating detected in the foyer. When empty, the average sound level produced essentially by equipment is 58 dB(A), whereas in busy periods it is around 71 dB(A). When there are fewer people in the early morning and late afternoon, this level dips a little below 70 dB(A).

This analysis by sensory modality helps us understand visitors’ perception scale and the socio-perceptive efficiency of the built environment. In the case of the Hall Napoleon, there is a divergence between visitors’ acoustic and visual experiences: acoustically, entering the pyramid corresponds to an immediate change in the acoustic milieu involving strong acoustic immersion and a reduction in hearing range; visually, however, the transition occurs slightly later, when moving towards the basement (change in perspective) and involves an asymmetry in conditions of visibility between the spectators located up above and the actors located down in the foyer below. This type of analysis needs to be pursued in two ways: first by looking at other sensory modalities (thermal, olfactory, etc.) in the same way; and secondly, by trying to show how the different phenomena identified by sensory modality fit together and work to create an overall ambience.


Notes on methodology

The commented walks method can be applied relatively broadly: e.g. to neighbourhoods, shopping malls, museums, railway stations, underground networks, transport hubs, underground public spaces and large urban projects. However, the need to move through a circuit does impose a restriction on the scale of the site studied. As such, it would be difficult to apply this methodology to extremely small territories (for example inside a home) or on a very large scale (an entire city). This type of field research is therefore particularly well suited to studying urban ambiences in situ.

This approach to urban space touches on a fundamental challenge in social sciences methodology: determining what is observable. Commented walks obviously draw on the reflexive capacity of human actors and on their ability to understand, describe and interpret the situations in which they find themselves. This use of discursive conscience is surely full of lessons, but it needs to be combined with the practical conscience of the agents involved. In particular, we can ask ourselves to what extent an ambience can be verbalised and actually concerns the field of language. How is it possible to put into words what is first and foremost a matter of physical sensation and immediate experience? This vital and difficult question will not be resolved here. We shall limit ourselves to a few summary pointers that will help clarify the scope of the debate. To begin, language is not a simple representation of the world, a tool amongst others to interpret reality or a simple instrument which allows us to communicate our experience. More fundamentally, language is what mankind uses to “control things” and “confront the world”[5] As such, we have posited that using speech to access ambiences allows us to understand how the latter are formed and shaped within and by language. In other words, the descriptions of perception obtained with the commented walks method primarily provided access to emerging sensory phenomena and to the configuration operations at work in perception. We can thus question the very possibility of verbalising immediately – here and now – what is felt or perceived in situ. Indeed, phenomenology has taught us that apperception and reflexive action are very difficult in the present. Becoming aware of one’s own activity is more conceivable after the fact, once an experience is over. From a methodological perspective, this issue was addressed in two ways. First, moving, travelling through different and contrasting ambiences allowed people to take a prospective and retrospective approach. The existence of thresholds and transitions encouraged comparison and putting immediate experiences into perspective. The comments did not refer solely to the present experience but also to the – different or expected – transitory phenomena which structure what has just occurred and which also allow us to anticipate what is to follow. In this case, the goal was not to extract oneself from an ambience completely, but rather to include one’s temporality and variations from within. Second, each walk was followed by a semi-structured interview which allowed the participant to go over what had just been experienced. Short-term memory was called upon to reinterpret and round out the initial comments. Far from being resolved once and for all, these questions will have to be the focus of future research and observation to confirm the validity of such an approach.

To conclude, commented walks quite clearly lead us away from everyday life situations. Although factoring movement into the research approach allows us to get closer to the ordinary activities of city dwellers, we still need to distinguish between the different scenarios that exist. The descriptions of perception in motion deserve to be analysed according to the passer-by’s course of action at the time of the experiment. We could as such point up how the sensory environment is shaped by the activity underway but even then we would still be a long way from a “natural situation” since the obligation to describe would continue to have an influence. That is why data gathered through “naturalistic” observation would allow us to put the final results into perspective. Conversely, the results obtained from commented walks raise the question of whether such naturalistic observation is actually possible. In analysing the sensory contexts of urban space, we demonstrated how they shape different types of visibility and audibility in public. If an ordinary passer-by can hear and see someone else more or less well depending on the sensory characteristics of a place, the same is true for ethnographers studying social behaviour. As such, the urban setting cannot simply be seen as a space in which behaviour occurs that merely needs to be recorded and analysed. What can we observe from in situ behaviour when passers-by appear overexposed, blurred in a poorly lit background or when they disappear from the frame altogether? How can conversations be recorded when background noise masks the voice or when the reverberation in a place limits the intelligibility of speech? The micro-ecology of public relationships is not merely the focus of theoretical thinking; it also questions the investigative methods upon which it is based. The public space issue thus confronts the researcher with the paradox of what is observable: describing situated actions based on observations which are themselves influenced by the site. Addressing this paradox is undoubtedly one of the major challenges of an observation methodology for urban ambiences.


Ackermann W., Conein, B., Guigues, C., Quéré, L., and Vidal, D. (Eds.) (1985). Décrire un impératif? Paris: CEMS-EHESS.

Arendt, H. (1961). Condition de l’homme moderne. Paris: Calmann-Lévy.

Augoyard, J.F., and Torgue, H. (Eds.) (2006) Sonic Experience: A Guide To Everyday Sounds. Montreal: McGill-Queen’s University Press.

Austin, J.L. (1964). Sense and Sensibilia. Oxford: Oxford University Press.

Barbaras, R. (1992). Motricité et phénoménalité chez le dernier Merleau-Ponty, in M. Richir and E. Tassen (Eds.),  Merleau-Ponty, phénoménologie et expériences (pp. 27-42). Grenoble: Jérôme Million.

Chelkoff, G., and Thibaud, J.P. (1993). L’espace public, modes sensibles. Les annales de la recherche urbaine, 57-58, 7-16.

Gadamer, H.G. (1976). Vérité et méthode. Paris: Le Seuil.

Gibson, J.J. (1986). The ecological approach to visual perception. Hillsdale: Lawrence Erlbaum Associates.

Goffman, E. (1963). Behavior in Public Places. New York: Free Press

Joseph, I. (1998) La ville sans qualités. Paris: Editions de l’Aube.

Leibniz, G.W. Nouveaux essais sur l’entendement humain (1966) [1704] Paris : Garnier-Flammarion, p. 38

Merleau-Ponty, M. (1968). The Visible and the Invisible. Evanston: Northwestern University Press.

Quéré, L. (1991). Qu’est-ce qu’un observable? In I. Joseph (Ed.), L’espace du public. Les compétences du citadin (pp. 36-40). Paris: Plan Urbain / Editions Recherches.

Quéré, L. (1992). Le tournant descriptif en sociologie. Current sociology, 40(1), 139-165.

Revue Enquête. anthropologie, histoire, sociologie. (1998). La description, 2.

Sacks, H. (1992). Lectures on Conversation. Oxford: Basil Blackwell.

Simmel, G. (1997) [1907] “The sociology of the senses” in D. Frisby & M. Featherstone (eds.) Simmel on Culture. London: SAGE Publications, pp. 109-120

Straus, E. (1935/1963). The Primary World of the Senses. New York: Free Press.

Thibaud, J.P. (Ed.) (2002). Regards en action. Ethnométhodologie des espaces publics. Grenoble: A la Croisée.

Whyte, W. (1988) City. Rediscovering the Center. New York: Doubleday

Jean-Paul Thibaud, sociologist and urban planner, is senior researcher at CNRS. Researcher at Cresson (Research Center on Sonic Space and the Urban Environment, UMR1563 Ambiances Architecturales et Urbaines). His field of research covers the theory of urban ambiances, ordinary perception in urban environment, sensory culture and ethnography of public places. He has directed the CRESSON research laboratory and is currently the scientific coordinator of the International Ambiances Network ( Jean-Paul Thibaud has published numerous papers on urban ambiances and has co-edited two books (among others) on this field of research.
email :
PDF version of this article


  1. See Simmel (1997); Goffman (1963); Whyte (1988); Joseph (1998) ^
  2. As Hans-Georg Gadamer (1976) has argued: “Not only is the world only a world to the extent that it is expressed in a language, but language itself only really exists by the fact that the world finds its presence within it” (p. 467, translated here). ^
  3. “Sound effects” and “frames of visibility” are central concepts underpinning this type of analysis. For the former, see Augoyard & Torgue 2006; for the latter see Chelkoff & Thibaud 1993. ^
  4. As an example, in terms of acoustics, we measured the sound levels (Leq), noise decay and reverberation time. With regard to lighting, we measured the luminance and illuminance. With regard to the thermal and aeraulic levels, we examined the air temperature, mean radiant temperature (wall temperature), air velocity and humidity. ^
  5. See Gadamer, op. cit. ^

2 comments for “Commented City Walks

Leave a Reply

Your email address will not be published. Required fields are marked *