4.1.1. On structure: Consecution versus integration

The audio descriptive guide is a text genre in which different discourse segments, also called "discourse moves," which may be descriptive, expository, argumentative and instructional, are combined to facilitate access to the museum (Soler Gallego 2016). In the audio descriptive guides included in this study, these different discourse moves are offered in either a consecutive or an integrated manner. At the consecutive extreme, found in the Museo Reina Sofía, the audio description of the artwork is recorded in a separate track, followed by instructions to access one or several tracks with expository and argumentative moves provided by museum experts and artists. At the integrative extreme, on the other hand, the audio descriptive, expository, and argumentative moves are included in the same track, either sequentially or in an intertwined fashion. This integrative approach is used in guides to both the museum's permanent collection, as in the Museo Nacional del Prado, and temporary exhibitions, such as the one created by Antenna International for the "Watteau The Drawings" exhibit at the Royal Academy of Arts in London or the one produced by Rouge Vivier for the "Icones Americaines" and "Velazquez" exhibitions at the Grand Palais. Interestingly, this approach has also been adopted for the audio description guide in two antithetical situations. One of them is when a special audio guide is created for a tactile exhibition located in a separate space of the museum, a solution adopted by the Museo Nacional del Prado and the Museo de Bellas Artes. The other is the creation of a universal design-based multimedia guide intended for adult visitors with all kinds of visual capabilities, an approach followed by the Open Art project in Poland (Szarkowska et al. 2016). A universal design approach was also implemented by GVAM for the Museo Sorolla, Museo Lázaro Galdiano, and Museo Julio Romero de Torres in collaboration with the CESyA (Spanish Center for Subtitling and Audio description) and later, for the Museo Carmen-Thyssen in collaboration with Aptent. However, GVAM approaches universal design in a different way. The goal of their approach is not to create one single audio guide to be used by visitors with all kinds of visual capabilities but to create two audio guides, one with audio description and one without it, which are as similar as possible in their contents. Their goal is to provide a similar experience with both guides, while offering more visual information for visitors who decide to use audio description. A consequence of adopting this universal design approach is that the specificity level understood as the level of conceptual detail and quantity of information in the audio description diminishes when compared to the majority of audio descriptions of visual artworks at museums and exhibitions included in the study. Between these two extremes, we find audio descriptive guides where minimal expository and argumentative information is included in the audio description track, and this is complemented with further information provided in a separate track. This middle ground approach is the one followed by the Museum of Modern Art, Tate Modern, and Solomon R. Guggenheim Museum.

Differences between museums at this discursive level seem to be related to two factors: (1) integrating the audio descriptive guide with existing audio or multimedia guides and (2) the overall design approach. In the Tate Modern and Museum of Modern Art, the audio descriptive guide was combined with an existing general guide for adult visitors so both audio guides can be accessed through the same device. In order to facilitate content navigation, specific instructions were included at the end of each audio description track to connect both guides. On the contrary, at the Museo Nacional del Prado the various audio and multimedia guides of the museum are not integrated but are accessed through different devices according to the type of visitor. The latter may seem a less inclusive approach, but it may be related to the fact that some museums prefer to use analog devices with a simple button pad for their audio descriptive guide because they find them more accessible than tactile devices. The following section deals with the second factor and more specifically, with a recent trend that advocates creating mobile guides which follow the principles of universal design so they are appropriate for adult visitors with all kinds of visual capabilities.

4.1.2. On content: One-for-all versus various-for-many

As mentioned earlier in this article, in the Open Art project's approach to art museum accessibility, audio description "is integrated in the general audio guide as part of the information provided to anybody using it" (ADLAB 29). Some years ago, a Canadian non-profit organization introduced this type of integrated or "intradiegetic" audio description in their accessible TV productions by training presenters and participants to intertwine their discourse with audio descriptions of relevant visual elements (Accessible Media). The use of this type of audio description has also been proposed for film (Thompson 2018) and applied to theatre (Udo et al. 2010; Whitfield and Fels 2013; Fryer and Cavallo 2018). A different definition of "integrated audio description" has moreover been advanced for theatre (Naraine, Mala, and Whitfield 2018), where it refers to the consideration of audio description from the commencement of the production process with the incorporation of the audio describer or a team member with the requisite skills in audio description:

Live integrated AD gives the director control over the development and execution of the AD, similar to his or her role in other aspects of audiovisual production, including, for example, costumes, props and sets… proponents of having AD being produced and delivered by the creative team suggest that there is potential for directors to exert creative control over the AD. Therefore, the describer's roles and responsibilities in each production are dependent upon how involved the director would like to be or can be in the creation, development and execution of an AD strategy… (Naraine, Mala, and Whitfield 115-116)

This integration and collaboration has been proposed as an essential aspect of "accessible filmmaking," where the aim is "to integrate audio visual translation and accessibility as part of the filmmaking process through collaboration between translators and the creative team of the film" (Romero 218). The question then arises as to whether audio description should be visible and differentiated as a resource specially intended for blind and partially blind people, or invisible and integrated with non-descriptive segments in a discourse intended for all visitors. When the latter is chosen, the specificity of the description provided is reduced, as illustrated by the Open Art project guide. This seems to be based on the assumption that highly specific audio descriptions are unnecessary and even inadequate for many visitors, and therefore, a balance between the needs of different profiles is sought by these means. However, to the best of our knowledge, this assumption has not been sufficiently corroborated by reception studies, and so the opposite could also be assumed. Why not create universally accessible audio guides with more specific audio descriptions for all visitors? At the same time, based on the results of the corpus analysis carried out in this study, there is considerable variation regarding specificity levels within the group of audio descriptive guides specifically intended for blind and partially blind visitors, with descriptions ranging from 100 to 800 words approximately. This seems to indicate that there is little agreement among practitioners on what the most adequate specificity level is.

Even if specificity were thoroughly researched via reception studies, the results would probably be inconclusive within groups of people with similar visual capabilities, for their responses are dependent on their sociodemographic and psychological profile, which in turn determines the type of experience they prefer to have at a museum, be it social, emotional or intellectual. For instance, Georgina Kleege, who is a professor of creative writing, literature and disability studies and is blind, is critical of the objective, highly detailed ADs where information is provided systematically so that "the blind viewer is taken on an orderly tour around the image" as found in some museums like the New York Museum of Modern Art. She argues that this type of AD should be replaced with more subjective audio descriptions where the impressions, both physical and conceptual or ideational, triggered by the artwork, are prioritized and presented in an order that is closer to sighted peoples' visual experience of art, that is, starting with the most striking elements in the work (112-117).

Since the type and quantity of information preferred by visitors can vary widely, the best option seems to be to create multilayered mobile guides with separate tracks offering different types of information, which can be chosen by the visitor as needed and preferred. This is the approach advocated by Proctor for museum guides: "Since there is no such thing as one-size-fits-all interpretation, stops should be layered, and ideally enable access to a variety of types of content to satisfy different kinds of questions and learning styles" (2010). As suggested by the study carried out by RNIB and VocalEyes, the situation is not different for audio descriptive guides. In this study, blind and partially blind people evaluated the audio guides offered by museums and heritage sites in the United Kingdom and one of the conclusions was that "The ideal guide was seen as being an integrated guide that has information for blind and partially sighted visitors on a mainstream guide. This information should be layered so that those who wish to access additional description or orientation information can access it" (29). Since it is impossible to design a guide that is tailored to every possible visitor, contents are either designed with those with the least access in mind or the offer of contents is diversified and users are allowed to build their own experience.

4.1.3. On the senses: From verbal to multimodal

In some of the museums included in the study, the sole accessibility resource provided by the museum to blind and partially blind visitors is audio description. Under these circumstances, completely blind visitors perceive the artwork through its translation into an aural-verbal text, namely the audio description. For partially blind visitors, the perception of the artwork is audio-visual and therefore, multimodal or multisensory. Many museums complement the audio descriptions with a tactile element, be it the original exhibit or a replica, giving rise to a second type of multimodal experience of the artwork. The Petit Palais in Paris adds a fourth sense to the experience by inviting visitors to smell different odors related to the exhibit such as humidity, wood, flowers, and fruit, created by a specialized company. In a holistic and inclusive approach to museum audio description, Eardley et al. (2017) have turned to studies in psychology and neuroscience showing how multisensory perceptual experience improves cognitive processing, in order to advocate for multimodal exhibitions that combine touch and sound, both verbal and nonverbal. The main rationale behind this approach is that multisensory experiences enhance learning of factual information and the creation of an autobiographical memory, i.e. "a memory which feeds in to and enriches their notion of self and can linger throughout life" (202), for every kind of visitor.

With regard to the aural mode, some museums have experimented with different types of non-verbal information. The National Gallery in London, for instance, includes sound interpretations of paintings along with verbal contents in their audio guide (The National Gallery). A different kind of non-verbal sound is used by Claire Bartoli during the audio descriptive guided tours she leads at the Musée d'Art Contemporain du Val-de-Marne, in Paris. This blind writer and actress uses paralinguistic elements, namely her voice tone, volume and speed, to add meaning to the conceptual content conveyed in the verbal components of her audio descriptions of artworks (MAC VAL). She was also commissioned to create recorded audio descriptions for three exhibits of the Musée d'Art Moderne in Paris (Mamvp). In an interview conducted for this study, the coordinator of the education department of this museum said that Bartoli interviewed different experts at the museum and used the information collected to create the artwork descriptions. The same coordinator thought that this type of description could be useful for all types of visitors and stated that for Bartoli, it is important that audio descriptions include subjective impressions of the describer because it helps blind and partially sighted individuals create their own interpretation. In a similar fashion, Neves (2012), a Translation Studies researcher and audio describer, draws on ekphrastic poetry to propose more subjective and creative ADs. She proposes that ambiguous visual signs be translated into equally ambiguous verbal signs seeking to have cognitive effects in the receiver similar to those provoked by the visual message. Neves also collaborates with sound producers to create "sound paintings," where verbal signs are integrated with non-verbal sounds in order to facilitate the interpretation of the verbal element. One example of this is the audio descriptive guide for the temporary exhibition "Olha por mim," composed of tactile paintings created by plastic artist Mirtilo Gomes (Lopes). The integration of non-verbal sounds is also found in the descriptions for the "Art History Through Touch and Sound" educative materials developed by Lou Giansante for Art Education for the Blind, and those for the "Star Spangled Banner" exhibition at the National Museum of American History in Washington D.C., created by Joel Snyder.

An emphasis on tactile resources continues to be present in numerous museums. According to Hayhoe (2013), this is due to the assumption that touch is preferred by blind and partially blind people when accessing the visual world. Hayhoe argues that this assumption is related to the evolution of Western ideas on blindness since the 17^th century and ignores the fact that most 21^st century blind and partially blind individuals in countries like the United Kingdom are not congenitally blind and thus, have visual memory. He claims that, given the diversity of needs, a plurality of resources needs to be present at museums so that both the majority and minority groups within the community of blind and partially blind people can enjoy real access to these experiences. Meanwhile, the curator of the Anteros tactile museum Loretta Secchi emphasizes the importance of touch in improving the access of blind and partially blind people to visual art (Secchi 2013, 21), as "The refinement of tactility facilitates the conscious development of one's perceptive autonomy and the function of visual thinking [my translation]" (Secchi 67). This view coincides with the results of two studies carried out in the United States (Reich et al. 7) and Spain (Cabezas Gay 2017), which showed the importance of the tactile experience in facilitating the access of blind and partially blind people to museums.

4.1.4. On intermodal coherence: From description to instruction

When touch is used during an audio descriptive guided tour, such as those organized at the Musée d'Orsay and the Museo Sorolla, verbal instructions are normally given to help visitors explore the tactile elements. However, when it comes to the autonomous visit with an audio descriptive guide, some museums offer both audio descriptions and verbal instructions, while others only offer audio descriptions. Among the audio descriptive guides that offer verbal instructions to explore the tactile image, we find various categories. The audio descriptive guide for the replicas at the "Furniture Gallery" in the Victoria & Albert Museum, for instance, indicates the location of the different components of the exhibit, but does not include specific instructions geared towards a tactile exploration (Victoria & Albert Museum, "Audio descriptions"). Moving towards the instructional extreme, in the audio descriptive guide for the touch objects located in several galleries devoted to Europe, the sentence "You're welcome to touch it" is the only verbal instruction and connection to the tactile object that is found in the text (Victoria & Albert Museum, "Europe 1600-1815"). On a further level of instruction, the audio descriptive guide for the autonomous "Touch Tour" in the same museum offers detailed guidance on exploring objects in the sequence illustrated by the following excerpt: "You can first locate his high hairstyle or hat, square in form. If you follow his arms down from the shoulders you will find the wide sleeves and beneath the hands notice how the curved folds of his robe are precisely detailed. The Buddha's hands are in the position of meditation, one resting in the palm of the other. Have you noticed the round object he is holding in his hands? " (Victoria & Albert Museum, "V&A Touch Tour").

When interviewed, Höelle Corvest, accessibility coordinator at the Cité des Sciences et de l'Industrie in Paris stated that verbal instructions should be used to help explore tacti images. Her rationale was based on the high difficulty in understanding a 2-D tactile representation of a two- or three-dimensional artwork and the low tactile literacy levels for this type of resource. Corvest's view was shared by two French companies specializing in tactile graphics for museums, La Ville Braille 2 and Tactile Studio 3 . The interviews revealed that all these experts coincided in affirming that the experiences of blind and partially blind of tactile resources vary depending on sociodemographic factors including their visual capabilities, educational level, and previous experience in visiting museums and using tactile images. Tactile Studio's representative highlighted the importance of assessing visitors' experience through reception studies with large numbers of participants. According to La Ville Braille, given the heterogeneity of users of tactile resources, the results of reception studies cannot be generalized to create a set of guidelines or good practices. This company advocated for offering tactile reading workshops to enhance the usefulness of tactile resources, and this is precisely what the Cité des Sciences et de l'Industrie has been doing. They organize tactile literacy workshops for both blind and partially blind people and individuals working on accessibility at museums and at tactile graphics companies.

4.1.5. On the voice: Human versus machine

Based on the results of this study, the majority of the recorded audio descriptions in art museums use human voices. Only a number of museums in Spain, namely the Museo Sorolla, Museo Julio Romero de Torres, Museo Lázaro Galdiano, and Museo Carmen-Thyssen, use an app developed by a company called GVAM (Accessible Virtual Multimedia Guides for Museums) which provides audio descriptions produced with a text-to-speech software. An interview with a company representative revealed that this technology is used because it allows companies and institutions to edit and add contents at any time, simplifies the process, and helps reduce production costs. Aptent, the company in charge of creating the audio description scripts for GVAM, also supports the use of text-to-speech software. In fact, they used this technology for the theatre audio description app they developed, which so far has been favorably reviewed by its users. Based on an interview with one of their audio describers, their rationale for using it coincides with that of GVAM. Moreover, they are of the opinion that text-to-speech voices make for greater homogeneity in the vocal qualities of the audio description, since the same human voice can sound differently at different times but text-to-speech voices are more uniform. In contrast, an audio describer and educator for the Museo Reina Sofía and the coordinator of audio description projects for the Grand Palais at the organization Rouge Vivier considered human voices to be more appropriate. According to the latter, human voices are more empathetic and personal and can be used to reinforce the message conveyed through words. Another audio describer and educator for the Museo Reina Sofía believed that the decision by some museums to use human voices is simply based on the fact that this is what other museums do, rather than on studies weighing the relative merits of all available options. On this topic, the education and accessibility coordinator at the Museo Nacional del Prado stated that they used human voices following the recommendation of the organization in charge of creating the audio descriptions, namely the ONCE (National Organization of the Blind in Spain).

To my knowledge, no existing research compares the use of text-to-speech versus human voice for audio description in the museum context. However, in the past decade a number of studies have looked into this variable for video audio description. Walczak and Fryer (2018) offer an overview of these studies from 2011 to 2015 and conclude that human or natural voices were preferred by participants for both fiction and non-fiction films, but many of them were willing to accept text-to-speech voices. Walczak and Fryer also refer to two other studies that showed that human audio description might be more appropriate for entertainment videos and especially for drama, while text-to-speech audio description would be recommendable for informative texts, such as instructional videos and documentaries. This is in line with the results of their own study, which showed that human audio description prompted significantly higher levels of presence for drama, while no statistically significant differences were found for the documentary genre. These results also coincide with those of a study of different voicing styles carried out by Iglesias Fernández et al. (2015), which showed that more interpretative audio descriptions improved receivers' experience of a dramatic feature film

4.2.1. On space: Inclusive versus exclusive

Kleege argues, in reference to the audio descriptive guided tours organized by some museums, that "specialized programs segregate blind visitors, when presumably the same methods could be instructive for anyone" (119). Some museums have opted for conducting this type of tours during regular opening hours. This is the case of The Metropolitan Museum of Art in New York and the Museo Thyssen-Bornemisza in Madrid. The Museum of Modern Art and Solomon R. Guggenheim, on the contrary, prefer to conduct this type of tour when the museum is closed to the general public. According to an accessibility coordinator of the Whitney Museum of American Art who was interviewed for this study, the museum uses this strategy during the touch tours for conservation reasons and to offer a quiet environment in which the auditory information can be perceived more easily. The National Gallery opted for splitting the museum visit into two stages. First, the group is taken to an educational space where the guide describes a painting and conducts the visual analysis of full-color, poster-size photographs of the work and enlarged photographs of details. Then, the same group is guided through the galleries for a visual experience of the artwork previously explored (The National Gallery). The education department at the Petit Palais also prefers this two-phase design. However, here visitors go first to the galleries and then to the educational space, where tactile reproductions of the paintings are explored.

With regard to audio descriptive guides, the large tactile reproduction of the L'atelier du peintre at the Musée d'Orsay allows autonomous visitors to listen to the audio description, explore the tactile representation, and look at the artwork simultaneously. This solution appears to be feasible when applied to a few isolated exhibits, but applying it to a greater selection of works would probably pose several problems for exhibition designers and curators because of the space it requires. Presuming that a museum visit requires considerable effort for blind and partially blind visitors, one might wonder how many of them would go to a museum to explore just one or two tactile images, replicas or even originals. The Andy Warhol Museum has incorporated a greater number of tactile reproductions into their exhibition space. This probably makes it more worthwhile for users of these resources to visit the museum. The use of smaller scale tactile images could limit the effectiveness of this resource in helping visitors understand the artwork. However, it facilitates their installation in the gallery space and thus helps visitors to integrate the different types of information (tactile, aural, visual) into a more coherent multimodal mental image. This inclusive design, where tactile images are located inside the galleries where the original artworks are exhibited, contrasts with separate spaces. At the Centre Pompidou, for instance, a series of tactile images created by Alain Mikli were located outside the exhibition rooms. Analogously, at the Museo Nacional del Prado and the Museo de Bellas Artes, the tactile images and the audio descriptive guides created by Estudios Durero 4 were installed in a separate gallery, similar to the Louvre's tactile gallery.

4.2.2. On the role of blind and partially blind people: From consultants to creators

In the museums under study, the authors of the descriptions for the audio descriptive guides are mostly external professionals, while those in charge of designing and leading audio descriptive guided tours are in most cases educators working for the museum. In both cases, most external audio describers and museum educators are not blind and partially blind people, although they collaborate with blind and partially blind consultants to develop accessibility resources. In a similar way, researchers in the field of audio description within Translation Studies, including me, are not to the best of my knowledge blind and partially blind people, and the role of blind and partially blind people is that of participants in studies of audio description quality and reception. However, at two museums included in the study, blind and partially blind people do actively participate in the creation of these resources. The Tate Britain, Musée d'Art Contemporain du Val-de-Marne, and Museo Reina Sofía are the only museums in our corpus where the audio descriptive guided tour is led or co-led by a blind person. Lisa Squirrel guides audio descriptive tours at the Tate Britain in London (BBC; Tate) and she is blind. Claire Bartoli, a writer who is also blind, does the same at the Musée d'Art Contemporain du Val-de-Marne, in collaboration with the curators and educators of the museum. At the Museo Reina Sofía, a sighted educator trained in accessibility by the ONCE organization audio describes the artworks, and a writer who is blind guides the tactile exploration of the works. As far as recorded audio description is concerned, Claire Bartoli has contributed to the audio descriptive guide of the Musée d'Art Moderne de la Ville de Paris. Additionally, the describer in charge of the audio descriptive guides for the Musée de l'Orangerie, the Musée d'Orsay, and the temporary exhibitions at the Grand Palais, has educational and professional experience in accessibility and is blind as well.

Museum accessibility benefits from interdisciplinary collaboration, and it is imperative that blind and partially blind people be an active part of this endeavor. To facilitate this, the development of accessibility plans for museums should be organized as participatory action research projects where a diverse group of researchers, practitioners, and users collaborate. The active role of beneficiaries in research on disability is at the center of the Emancipatory Disability Research (Hollins 2010). This is a theory-methodology that emphasizes the great value of their expert opinions, derived from their personal experience:

This new paradigm aims to give disabled people control over the research agenda, seeks to benefit those involved in the research process and ensure that outputs are accountable to disabled people, in the way their views and experiences are represented. This research philosophy marks a significant shift in thinking and aims to tackle fundamental power inequalities within the research process itself. In this new model, the role of the researcher as an expert is challenged as disabled people's embodied knowledge about their impairments is given an equal footing to the researcher's knowledge. One of the central aims of emancipatory research is that the process should be used as a tool to change society. (Hollins 228)

Among AD researchers, Di Giovanni (162) advocates for and implements a more (as compared to existing practices) emancipatory approach to the creation of opera audio description. She and other researchers and opera audio describers set up a participatory laboratory to write, revise, and record the audio description for the Carmen, la stella del circo di Siviglia opera performances with blind and sighted children and teenagers. In the first of two sessions, a draft audio description script prepared by the group of researchers and audio describers was used as a basis for discussion with the young participants. In the second session, the team recorded the audio descriptions. This is a positive advancement. However, it is the type of formative evaluation in which the audio description scripts and/or recordings are reviewed by blind consultants prior to making them available to the public that is present to some extent in many of the museums studied, including the projects carried out by Kaleidoscope 5 non-profit organization. Positive as it is for the design and implementation of an emancipatory model of museum accessibility research and development, this type of consultancy-based participation only reaches the second of the four levels of access of disabled people to "opportunities to shape institutional policy and practice" as defined by Hollins (235). These four levels constitute necessary steps for museums to build relationships with disabled people that truly reflect the principles of emancipatory research. They are the following:

1. Disabled people gain access to the museum, but decisions about access are made on behalf of them and without consultation.
2. The museum consults with disabled people on identified issues.
3. Disabled people are long term consultants, and consultation is a two-way dialogue that explores issues that are important to both parties.
4. Disabled people gain access to the decision-making process at the most senior level in the museum hierarchy and are able to directly influence decisions.

On the second level, relationships created between the museum and people with disabilities are short term and dissolve after tackling the issue at hand (Hollins 236). This is the approach I mainly identified in the exhibitions and museums in the corpus. The fourth and highest level of emancipation has been achieved in museums like The Metropolitan Museum of Art and the Cité des Sciences et de l'Industrie, which have employed blind accessibility coordinators. When it comes to the design, creation and implementation of accessibility resources, such as verbal description and tactile materials, blind and partially blind people should not be relegated to the role of consultants but can also lead and collaborate in the design and implementation of these resources as experts.