Tactile cartography in the digital age: A review and research agenda

I Introduction

At the time that Perkins (2002) wrote the last major literature review of tactile mapping research, the field was on the precipice of a technological sea change. Digital mapping’s influence was becoming unavoidable, even prompting Perkins to ask whether physical tactile maps are ‘still relevant in the digital and multimedia age’ (2002). But while he noted that many of tactile mapping’s concerns overlapped with those of cartography as a whole, the former was still regarded as ‘marginal’ to the latter, lacking the advantages of industrial-scale production, standardized design practices, and tactile-compatible data sets. Perkins speculated that new digital technologies may be able to address some of these imbalances.

If 2002 was the year when we gazed at the digital horizon of tactile mapping, then 2021 may be a good time to assess exactly what is on the other side, how we got here, and where we need to go. On the whole, tactile mapping remains a ‘specialist’ field, with many papers noting that tactile maps are still expensive, hard to come by, difficult to make (Brittell et al., 2018; Ghodke et al., 2019; Stangl et al., 2019), and inherently unable to match the resolution of data that is possible with a visual map, even with the affordances granted by digital, interactive technologies (‘Section 508’, 2014; Weimer, 2017). Perhaps these are problems that will never quite be solved. But is it possible that they do not need to be solved? Have the standards for tactile cartography been miscalculated? What does the practice look like if it is taken on its own terms? Who exactly will be making these assessments, conducting this practice, and using the maps? Historically, the primary audience for tactile maps has been people who are blind or visually impaired (hereafter B/VI), but some studies have pointed to the benefits of tactile maps for people who are sighted as well (i.e. Gual et al., 2014), expanding the audience considerably. What implications do these findings have for the field?

In beginning to answer these questions, I will pick up where Perkins left off, reviewing work in tactile mapping over the past 19 years and adding to contributions from other reviews. I build on work by Tatham (2003) who offers a brief research agenda, Rowell (2007) who focuses on location-based services, Koch (2011) who gives a brief overview of technological developments, Brock et al. (2013) who focuses on orientation and mobility mapping, and Wabiński and Mościcka (2019) who focus on automatic tactile map generation. The following review will cover these topics as they pertain specifically to tactile maps rather than related media. In that vein, I will also be limiting my scope of analysis to tactile maps rather than haptic maps more generally. The International Standards Organization (ISO) distinguishes between the two within the realm of human–system interaction, with ‘tactile’ referring to ‘mechanical stimulation of the skin’ and ‘haptic’ encompassing ‘all haptic sensations’ including tactile stimulation but also kinesthetic and force-feedback stimulation (ISO, 2011). Maps that are primarily tactile but incorporate haptic elements will be included in this review, as will research that is concerned with tactile graphics more generally but includes maps as a subcategory. However, the implications or conclusions of the research reviewed here are always directly applicable to tactile cartography.

The article’s structure is as follows. First, I review research on data acquisition and organization, including the conversion of visual map data to tactile map data, and the use of open spatial data sets in tactile mapping. Second, I look at research on map production, including methods for printing, publishing, or otherwise distributing tactile maps, with a focus on multimedia and interactive platforms. Third, I discuss research about tactile map design; while this has historically been a popular topic in research, participatory methods have begun to change exactly how that research is conducted. Fourth is a section on tactile map use, in which I will focus on tactile maps in situ, that is, how tactile maps fit into the lives of their users outside of experimental contexts. Fifth and finally, I synthesize the aforementioned research to propose an agenda for future tactile mapping research based on the four categories outlined here. This agenda will outline avenues of further inquiry that are not only productive in terms of scholarship but more importantly also attentive to the lived experiences of stakeholders.

II Data: Transcription and Translation

The standardization of tactile mapping practices has been on the to-do lists of many researchers for decades. While certain best practices have been established for tactile graphics more generally (‘Guidelines’, 2010), tactile mapping standards have never quite made it past the proposal stage. Rowell and Ungar (2003a) offer a proposal for a database of tactile symbols, but it would turn out to be one of the last papers on the subject. Coming from the opposite angle, Wabiński and Mościcka (2019) give a comprehensive overview of spatial database research that can be extended to tactile mapping, but the list of those developments is still missing a dedicated, public repository of tactile-ready data. For tactile maps specifically, research on ‘standardization’ has largely shifted away from a focus on standardizing sets of symbols or layouts and toward universally applicable processes for quickly creating tactile maps, often under the heading of ‘automation’. In other words, the process rather than the product is now the objective of standardization efforts (as much as any mapping process can be standardized). In this article, I categorize automation efforts in terms of translation and transcription. While used more or less interchangeably in tactile graphics literature (see Chen and Takagi, 2013 and Tanaka, 2019, for opposing uses of the term ‘translation’), I define transcription here to refer to the conversion of preexisting visual maps into tactile maps and translation to the tactile mapping of ‘raw’, or unmapped, geospatial data. In other words, translation is an act of synthesis, while transcription is an act of duplication, but each process presents its own unique challenges.

The transcription approach to tactile mapping is not as strongly represented as translation, likely due in large part to licensing and copyright issues associated with previously made maps. Transcription methods may include generating tactile maps from Google Maps or MapQuest tiles (Moustakas et al., 2007; Wang et al., 2009), hand-drawn paper maps (Takagi and Chen, 2014), or PDFs and JPEGs of building floorplans (O’Sullivan et al., 2015; Trinh and Manduchi, 2020). While the creation of tactile graphics almost always necessitates creative interpretation given that tactile media requires highly generalized features, a potential benefit of using existing maps as source material is that the source maps largely circumscribe the range of features and their arrangement in the output map. But of course, the proposition of transcription is becoming more onerous as more maps become interactive (which would pose some interesting challenges regarding queryable features), and issues associated with rapid zooming in and out are impossible to address without refreshable braille display technology (see Leonardis et al., 2018, for an overview of recent developments on refreshable displays). However, this research could also offer potential avenues for developing ‘rapid transcription’ systems that could quickly produce tactile graphics in the absence of tactile graphics design tools. Takagi and Chen’s (2014) system for producing tactile graphics from hand-drawn maps, for instance, presents obvious applications for the rapid dissemination of information in a tactile format, such as during a disaster response scenario. Historical and unique or limited-run physical maps can also be digitized and made tactile using these systems.

The other approach, translation, entails creating tactile maps directly from unmapped geospatial data (i.e. translating from data to map), such as TIGER® line data (Miele et al., 2006), LIDAR data (Schwarzbach et al., 2012), depth camera imagery (Katzschmann et al., 2018; Velázquez et al., 2006), data from OpenStreetMap (OSM), which is especially popular (Ducasse et al., 2015; Götzelmann and Eichler, 2016; Hänßgen et al., 2016; Poppinga et al., 2011; Rifat et al., 2011; Taylor et al., 2016; Watanabe et al., 2014), or using whatever data are available from public GIS data portals (Červenka et al., 2016; Štampach and Mulíčková, 2016). The data source used by each project influences the type of map that is created, as one might expect (e.g. LIDAR data are more often used for large-scale maps of buildings, whereas OSM data are used for neighborhood-scale maps), but there also appears to be a connection between data type and medium; dynamic, interactive tactile displays tend to be used for very large-scale spatial data such as building interiors, whereas static tactile media are more often used for smaller-scale maps. Experiments that swap these two media could generate interesting results, as static large-scale maps may prove to be more useful for certain tasks or processes than dynamic large-scale maps, for example.

As geospatial data are increasingly more detailed and become easier to obtain (Rice et al., 2013, describe a crowdsourcing approach to data collection), translation may prove to be a more commonly employed approach than transcription, in terms of both research and practical applications. Translation allows for a great deal more flexibility in terms of what is actually included on the map as well as the technology that could be used to make the map, as seen in Pandey et al. (2020) who describe research on developing a system for allowing people who are B/VI to draw their own tactile maps using a stylus and tactile film. However, as I mention above, transcription systems that help automate tactile map production could prove to be vitally important in certain situations.

Regardless of the approach taken, it may be illuminating for researchers and cartographers to consider why they are choosing to transcribe or to translate the data they are working with. Are the maps made to accommodate people who are B/VI but serve primarily sighted users? Or are they made for or even by people who are B/VI? These questions can, in turn, help us critically examine data production practices as well. Who is data being collected for? Will data for a given project be relevant to or usable by people who are B/VI? Certain geospatial data, for example, the locations of curb cuts, tactile paving, and accessible you-are-here maps could be beneficial to people who are B/VI (Kitchin, 2002) but may not be routinely collected in a geospatial survey of a city. At the same time, data that are routinely collected without being augmented by other data points may be less useful, such as road centerlines in the absence of sidewalk data or building footprints without entrance/exit locations (these issues are taken up by the framework of Universal Design (UD), discussed further in section IV). In most cases, available resources and public policy regarding information accessibility will dictate which approach is taken – that is, whether data can be collected in the first place, how it is collected, the quality of the data, the methods by which it gets distributed, what sorts of regulations and institutions oversee the process, and so on. But determining the character of one’s workflow may be able to help answer those questions.

III Production: Multimodality

Arguably the most visible developments in tactile mapping over the past 19 years have been those regarding the media used to create and distribute tactile maps. While embossed paper remains a popular medium (Lobben, 2005), technological developments, including the proliferation of mobile computing (Poppinga et al., 2011), widespread Internet access (Ivanchev et al., 2014), and rapid prototyping (Brittell et al., 2018; Taylor et al., 2016) have had significant impacts on the landscape of tactile mapping research. These developments have had a democratizing effect on research, enabling a greater volume of experiments to be conducted by a greater diversity of scholars (especially those who may not have previously had access to the necessary resources 19 years ago). But they are also generative of two important recurring themes in tactile map research: participation, which I will discuss in section IV, and multimodality.

Multimodal tactile maps are maps and mapping systems that incorporate multiple sensory interactions and input/output modalities (that is to say, they provide multiple options and tools for exploring spatial data), often two or more different modalities at once. This representation of spatial data in multiple parallel formats is described by Golledge et al. (2006) as ‘redundancy and augmentation’. Augmentation can take the form of sound, vibration, force feedback, temperature, pressure, texture (Griffin, 2001), and even scent (Brule et al., 2016). Golledge et al. (2004) also state that multimodal interfaces do not just provide alternative ways of accessing information for people who are B/VI, but also that they increase comprehension of that information. Importantly, they explain that this principle also applies to those who are not B/VI, meaning that collaborative tasks may benefit from using multimodal tactile maps. A burgeoning methodological approach may also lend insight into the virtues of different modalities: tactile psychophysics uses neurophysiological data to augment and inform perceptual stimulus data. As noted by Jehoel et al. (2006, 2009), psychophysics has yet to be satisfactorily linked to tactile map design decisions; however, it may provide guidance on establishing certain guidelines, such as the minimum necessary elevation of features, or which printing media improve readability. Being able to quantify tactile perception beyond the point where useful descriptive feedback becomes difficult for users to give – whether or not the user is B/VI – could have implications for just about every stage of the research process, from methodology to usability.

Multimodal tactile interfaces offer, as one would expect, a great deal of flexibility when it comes to representation of spatial data. A widely cited paper by Miele et al. (2006) introduces one of the first mapping platforms that used text-to-speech (TTS) in place of braille labeling, an approach adopted by many following studies. Additionally, mapping platforms that also respond to audio input are becoming more feasible with improved speech recognition technology (Abd Hamid and Edwards, 2013; Barbosa and Sá, 2020; Cavazos Quero et al., 2019; Reinders et al., 2020). Mobile computing is also frequently enrolled in multimodal technologies, such as in Yatani et al. (2012), Matsuo et al. (2020), and Giudice et al. (2020) who incorporate tactual, auditory, and vibration feedback in mobile wayfinding applications, or Senette et al. (2013) who lay a microcapsule tactile map over a mobile device with an app installed that can recognize it. Kane et al. (2013) discuss a similar set of experiments, except in their case, they use 3D-printed touchplates to guide users’ fingers while using a touchpad.

3D printing has proven to be a popular system for experimentation, given its capacity for easily creating volumetric tactile media. The notion of a 3D tactile map has itself been around for decades, often manifesting via a process called thermoforming or vacuum forming in which a heated sheet of plastic is draped over a mold to create a relief of whatever area is being represented. Rowell and Ungar (2003b) found that thermoform maps were popular (at the time of the article’s writing) due to their durability and capacity for representing volumetric data. However, Rice et al. (2005) and Gual et al. (2015) both note that thermoform graphics are often not portable given that they are rigid and unable to be folded without introducing tactile variation by way of a crease, and that creating a mold is expensive and difficult, limiting potential thermoform cartographers to only those who specialize in the practice. Other materials have been used to make 3D maps as well, such as wood and metal (Kent, 2019), but 3D printing technology has continued to become a more viable option, not just in terms of what can be symbolized (which will be discussed in the next section) but also via its capacity for rapidly prototyping maps, especially those with interactive features.

Interactive tactile maps in general may improve the acquisition of spatial knowledge and general user satisfaction (Brock et al., 2015; Brule et al., 2016; Giudice et al., 2020) as well as improve performance with specific tasks such as wayfinding (Palivcová et al., 2020; Toyoda et al., 2020). 3D printed maps also provide a relatively affordable method for producing simple, custom interactive media (Giraud et al., 2017; Gual et al., 2014; Rener, 2017). Additionally, 3D printed maps may prove to be more intrinsically comprehensible by people who are B/VI. Holloway et al. (2018) conjecture that 3D printed maps reduce the cognitive load for map users by reducing the need for graphical tools that help illustrate an abstracted 2D landscape – such as side views or occlusion – allowing for a more ‘purely’ tactile experience; however, they note that it is unclear as to whether these benefits extend to maps specifically. 3D-printed touchplates that attach to tablet computers have been a popular research topic (Götzelmann, 2018; O’Modhrain et al., 2015), but the user experience has not been uniformly positive. Kane et al. (2013) observed that users in their study were reluctant to learn how to operate a new device, although the users noted that it may be useful as a training device for people who have recently become B/VI. The authors also observed that users were not entirely happy with the tactile experience of their touchplates, a result also noted in Brock et al. (2010) and Taylor et al. (2016). However, audio integration with touchplates has shown promise, such as those that use TTS for reading labels and directions (Götzelmann, 2016; Shi et al., 2020; Simonnet et al., 2019).

While multimodality is often enrolled in helping B/VI users explore information, the implications are in fact much broader. Brule et al. (2016) conducted a survey on accessible maps in an educational institute for children with various disabilities. They found that that children need to be able to explore maps and discover information on their own, that caretakers need to be able to easily create new material, and that children need to be able to collaborate with their caretakers as well as with children who have other disabilities. Multimodal tactile maps may be able to address each of these needs, perhaps not as a single system but certainly as an approach. Barvir et al. (2018) discuss tactile maps that include text in both braille and Latin scripts, allowing both sighted and B/VI people to use the same map together. Giraud et al. (2020) examine a series of workshops in which both sighted and B/VI residents of a rural village worked together to design an interactive tactile map of the area. The authors noted that the actual design process itself (i.e. meeting in a room, working with physical materials) provided as much insight into the evaluation of designs as the formal evaluation itself. In all these cases, multimodality enables augmentation and repetition of information, but it also enables more balanced collaboration between people with differing sensory abilities, which brings us to the next recurring theme in contemporary tactile mapping literature: participation.

IV Design: Participation

While a given visual map from the mid-1800s may look strikingly different from one created today, the design of tactile maps has not changed much in their 150 or so years of existence. Credit for the atlas is a matter of some debate, but I will defer to the publisher’s sentiments and give precedence to its printer, Stephen P Ruggles, in the following citation: Ruggles and Howe (1837). One of the first tactile atlases ever published includes tactile maps that closely resemble the designs we use today – geographic features are represented simply and sparsely with only the most salient features being included on the map with ample whitespace between them. Labels on the map take the form of raised Latin letters, a practice which has since been replaced with braille labelling, but most other methods of symbolization have stayed close to their original designs. Instead, the biggest change seen in the last 19 years has been the process of designing these maps. Specifically, participatory design broadly defined has become an increasingly popular methodological approach, especially over the past decade.

Participatory research has existed as a methodology for many years in social science (Maguire, 1987) and design research (Carroll, 1996). Participatory disability research aims to address what has been a historically exploitative power imbalance between nondisabled researchers and disabled participants (Barnes and Mercer, 1997; Oliver, 1992). By designing the research protocol to be directed in part by the participants, the process thereby (theoretically) better represents the needs and desires of the participants rather than just those of the researcher (Kitchin, 2001, 2002). Here, I draw a line between participatory research and participatory design – while the terms are sometimes used interchangeably, the former will refer here to decisions concerning research protocol, while the latter will concern the design of an artefact, but not necessarily experiments and interviews, and so on (see Albouys-Perrois et al., 2018, for a brief overview of literature on adapting participatory design practices to people who are B/VI). In drawing this distinction, participatory design is more closely aligned with other progressive design approaches such as codesign (Coughlan et al., 2020; Yusim, 2019), user-centered design (Bardot et al., 2017; Rassmus-Gröhn, 2008) or inclusive design (Chamberlain and Dieng, 2011; de Almeida, 2014). This is not simply a rhetorical move, as important questions have been raised about the degree to which B/VI participants actually have any control over the research protocol and how they are represented in the final research output.

Vermeersch and Heylighen (2019) explore questions of control in research by describing architectural codesign sessions that incorporated building models arranged on tactile maps. They note that when B/VI participants (referred to in the paper as ‘user/experts’) were exploring the tactile media, ‘the ownership of the meeting’s narrative shifted from the architects to the user/expert’, because the architects had to negotiate a more haptic vernacular rather than one that was primarily visual. But this was not extended to the remainder of the project: ‘such shift was not observed for the design itself, as the setting…made [the user/expert] reluctant to make any design move’, that is, the user/expert deferred to the expertise of the architects regarding design of the models themselves even though the user/expert felt confident in providing their impressions of those designs (Vermeersch and Heylighen, 2019). There is clearly a dynamic between power and knowledge being manifested in this study; one that may in fact become amplified by the trappings of empirical science (Spinuzzi, 2005). Following this line of critique presents opportunities for impactful future research, although it must address not just the design sessions but the entire life cycle of a study, including and beyond publication.

There is also potential for investigation into how participatory research intersects with other types of inclusive research methodologies. For instance, Bernardi and Kowaltowski (2010) and Kowaltowski et al. (2015) both employ tactile maps as part of architectural design workshops involving people with disabilities. The authors situate their research within the principles of UD, which aims to produce designs that are accessible to as many people as possible thus including people with disabilities (Hamraie, 2013) rather than designing specifically or exclusively for people with disabilities. This raises an important point of clarification: UD is not inherently participatory, as a team of entirely nondisabled designers could technically produce something that follows the principles of UD. The work done by Bernardi and Kowaltowski (2010) was participatory because it was participatory, not because it employed UD. At the same time, participatory design research is not always universal. There is occasional slippage between these terms, but their strict definitions (as strict as they can be) have important implications not only for research protocols but potentially data collection and storage as well. After all, a tactile map may be created using ostensibly participatory protocols, but the data format or storage media used for the map could easily be inaccessible to people who are B/VI, raising questions about the actual degree of participation in the research. These circumstances, while perhaps not inherently problematic, illustrate the need to think holistically when considering participatory methods.

Participatory methods, while they may profoundly affect the process and outcomes of research, are often uncomplicated in their execution. Brock et al. (2013) and Ghodke et al. (2019) both employed participatory design in their research, which took the form of brainstorming sessions. While a relatively straightforward addition to the study protocol, the brainstorming sessions allow the authors to distinguish their research from a nonparticipatory study in which feedback would be gathered strictly through observation in experiments. Albouys-Perrois et al. (2018) detail how brainstorming can be implemented in research as demonstrated through their study on creating a multimodal augmented reality tactile map. The authors state that a longitudinal series of meetings generated more useful insights than just one meeting, which is consistent with the findings of Ghodke et al. (2019) and may seem like a self-evident conclusion, but multiple meetings are not always employed. Albouys-Perrois et al. (2018) also used an iterative prototyping process and evaluation by way of a workshop, in addition to observation, interviews, and structured evaluative experiments. These studies all demonstrate the range of avenues that can be taken to employ either participatory design or participatory research.

V Use and Evaluation: Access and Operation

Not long after Perkins published his 2002 review, Aldrich et al. (2002) published an article proposing a ‘model of tactile graphicacy’ or a conceptual framework for understanding how people read and create tactile graphics. In the following previous scholarship on visual graphicacy (Poracsky et al., 1999), they position ‘tactile graphicacy’ as a cognitive extension of literacy, and the paper in question focuses on how it is developed by people who are B/VI. So now that we are fully situated within the era of digital mapping, how do those technologies affect ‘tactile graphicacy’ or spatial cognition in the context of tactile mapping? Can we answer Perkins’ (2002) question about whether digital mapping technology will supersede physical tactile maps?

At best, the answer seems to be unclear. We do know that being B/VI has no bearing on one’s ability to cognitively process spatial information or perform spatial tasks (Caddeo et al., 2006; Lahav and Mioduser, 2008). That said, reading a tactile map is a much more cognitively intensive task than reading a visual map due to the need for constant referencing and orientation via touch (Ducasse et al., 2018; Perdue and Lobben, 2016), a task made even more difficult by the lack of a standardized symbol set. Digital tactile maps do have the capacity to reduce cognitive load by incorporating high-fidelity media such as audio information (Bardot et al., 2016; Götzelmann, 2018), reducing the need to memorize symbols and labels, and so on. Digital tactile maps can also encourage exploration (Zeng and Weber, 2016) and facilitate learning to create further tactile media (Rassmus-Gröhn et al., 2007).

There is also an important, steady beat of research on tactile map usability and cognition that does not necessarily involve digital or interactive maps, investigating topics including – but not limited to – symbolization (Lee, 2019; Rahardjo et al., 2019), estimating spatial relationships (Hagedorn, 2012; Simonnet et al., 2019), tactual perception (Bardot et al., 2014; Brock et al., 2012; Jansson and Monaci, 2003; Morash et al., 2014), representations of scale (Rastogi et al., 2013; Yayla, 2009), and the comparison of tactile maps to verbal descriptions (Habel et al., 2010; Koustriava et al., 2016). This research is certainly crucial for deepening our understanding of how tactile maps work and are used. However, it is nevertheless research that could have been carried out prior to 2002; that is, it does not represent any major evolution in usability research. This is by no means a shortcoming of any sort – research in this vein has simply been consistent.

In consideration of the foregoing, the most useful thematic through line of tactile mapping usage research might in fact be the absence of research on usage. That is to say, how tactile maps are actually used – not in an experimental setting but in the daily lives of people who are B/VI. Perkins and Gardiner (2003) briefly mentioned that there has been ‘little investigation into the “ecology of reading” [tactile maps]’, which unfortunately remains true today. However, this is not entirely a shortcoming that can be laid exclusively at the feet of researchers. Upon a close reading of the usability research that does exist, one will notice that users of tactile maps point to a consistent issue of access. This is not access in the sense of accessible design but rather the ability to acquire and use tactile maps on a regular basis and, ideally, in a range of contexts. In the words of Froehlich et al. (2019), ‘The data and underlying models [of tactile maps] are meaningless if a broad user base cannot access, interact with, and use them’. However, if someone can indeed access a tactile map, would they actually use it? And if so, how? I place these questions and their answers under the heading of operation, in the sense of directly manipulating a map as well as the constellation of associated procedures and considerations that one must take into account when using a map. Obviously, the physical process of using a tactile map differs from that of a visual one, but the times and situations during which a tactile map may be used (or not used) can differ as well.

In general, the literature suggests that B/VI study participants tend to show some amount of interest in – if not enthusiasm for – the maps that they are asked to use in studies. However, legibility of map design is often inconsistent between studies, and comprehension levels between participants in a single study can vary dramatically (see Soh and Smith-Jackson, 2004 and Retchless, 2014, for discussions on differences in perception of visual maps between individuals). While researchers often write tactile map legends and labels in braille, very few people who are B/VI are actually able to read or write braille, due in large part to the fact that the vast majority are people older than 50, and thus did not grow up learning it (Bourne et al., 2017), an issue also investigated by Gintner et al. (2019). The most cited statistic of braille-literate students is around 10 per cent, but this is contested – Graves (2018) provides a discussion on the issue. In the case of tactile maps specifically, Lobben and Lawrence (2012) note that several of their B/VI participants had never actually used tactile maps before the study due to a lack of access, resulting in diminished enthusiasm about using the maps, as well as somewhat increased difficulty in terms of comprehension. However, diminished enthusiasm seemed to follow from difficulty with using the maps; in other words, participants were interested in the maps themselves but anxious about actually using them, resulting in apprehension (Lobben and Lawrence, 2012).

Issues with access to maps and anxiety about using them are issues discussed extensively in Stangl et al. (2019). Participants in their survey consistently mentioned a chronic lack – or at least inconsistency – of easily accessible tactile media. This is especially true in environments that do not cater specifically to people who are B/VI, such as in a classroom with mostly sighted students (Rosenblum and Herzberg, 2015). On the topic of neighborhood tactile maps, one B/VI participant said that people ‘would recognize the streets that they knew and become emotional about getting access to this kind of information…And many of us stopped thinking that it was something we could have done’ (Stangl et al., 2019). At the same time, Stangl et al. (2019) also noted that there is an absence of spatial literacy education as it pertains to tactile maps; echoing Lobben and Lawrence (2012), there was interest and enthusiasm among the participants about tactile maps; however, it was accompanied by uncertainty as to exactly how the maps could be used. Golledge (2005) explains that tactile maps of local areas (as opposed to small-scale reference maps) are in high demand but are unavailable either due to their expense or lack of portability. Therefore, access to tactile maps is attenuated directly by a combination of funding, literacy, and legibility and indirectly by issues with the social dynamics arising from the use of tactile media.

Additionally, there is a general social stigma against the act of touching, which can lead to anxiety surrounding the use of tactile media. Stangl et al. (2019) note that the act of touching in certain contexts is frequently considered to be unhygienic and/or infantile, and as a result, people for whom touch is a primary means of engaging with the world are isolated both socially and intellectually when they are deprived of that option. The authors also explain that tactile media can often only be operated in a relatively conspicuous manner, thereby alienating its users, especially in environments conventionally associated with visual experiences (Stangl et al., 2019) – placards that read ‘Do not touch’ are ubiquitous in museums, but none that say ‘Do not look/hear/taste/smell’ (obviously touching art poses more of a risk of damage than looking at art does, but nevertheless, certain people are excluded from experiencing the art as a result; see Shaligram, 2019 and Rieger et al., 2019, for further discussion about museums and visual disability). Golledge (2005) and Giudice et al. (2020) both explain that tactile maps or map devices are often cumbersome and not designed to be portable and ostensibly therefore requiring conspicuous operation in public (it is worth noting the 15-year difference between these papers).

Although there are certain circumstances in which a tactile map could be useful but is unavailable or cannot be operated easily, there are also situations in which a tactile map might not actually be useful or would be useful in ways that differ from a visual map. In a study that focused on indoor tactile maps, B/VI participants noted that tactile maps may be potentially helpful for navigation, but that there is usually another person present in a given building who can help with navigation, reducing the need for a map in the first place (Trinh and Manduchi, 2020). Also, visual map users often consult a map before visiting a place physically, but a B/VI participant in the same study mentioned that she would actually prefer to use a map after she had explored a room, fitting her proprioceptive experience to the diagrammatic one presented to her rather than the other way around (Trinh and Manduchi, 2020). This approach is not commonly found in research, partly due to the structure of study protocols, but suggests that tactile maps may play a previously unconsidered role in developing a post hoc mental map of a space depending on the space in question. The good news is that this is a phenomenon that can be investigated in future research.

VI A Research Agenda for Tactile Mapping

I will conclude by assembling several specific lines of inquiry that tactile mapping research can take in the coming years. As demonstrated above, the advent of digital mapping has not done away with tactile mapping but rather has augmented it. People who use tactile media have repeatedly expressed their preference for some sort of haptic feedback over more abstract types of representation, whether in the form of tangible objects or vibratory cues. The way that this feedback is delivered can be varied – from embossed paper to 3D-printed models to a buzzing smartwatch – but nevertheless involves touch in some respect.

My research agenda for tactile mapping is framed by two premises. First, I reiterate Perkins’ (2002) conclusion that there remains much work to be done on the social dimensions of tactile mapping. The literature discussed here presents essential findings on the development and evaluation of various tactile mapping systems and technologies, and so the next major topic that needs to be addressed is how these maps are used in situ – how they both inform and are informed by the lives of their users. The second premise, following directly from the first, is that tactile mapping is a discipline in its own right – related to visual mapping, but with its own standards, expectations, practices, traditions, imaginaries, and trajectories. While visual cartography can be used as a point of contrast for tactile cartography, it should not necessarily be invoked as discipline to emulate but rather as a rich vein of inspiration. So, in that context, and using the categories listed above as a broad framework, I outline a tactile mapping research agenda for the digital age as follows.

One: Research directed by people who are B/VI

People who are B/VI should be involved at every stage of tactile mapping research; not only at the point of experimentation but also when generating research topics, designing their protocols, and when writing and publishing results. While the idea of participatory research may be an attractive one, the path to implementation is not always straightforward, nor should it be. Rieger et al. (2019), invoking Moser (2000), discuss inclusive spaces in the context of museums and offer some cautionary words about formalizing participation:

Inclusion in the museum is not easily produced, analyzed, or mapped…It cannot be done through the use of best practices or through prototyping and participatory user groups. [Inclusion] is about taking risks, becoming undone, admitting the limitations of people’s limited embodied experiences, and involving a breadth of human experiences (not just our own) to develop better ways of doing inclusion as an ongoing enactment (Rieger et al., 2019: 18).

I would like to take the arguments of Rieger et al. (2019) a bit farther and underscore the potential importance of experience in tactile cartography. Tactile maps are frequently a more embodied experience than visual maps, the affective triggers of visual cartography (Anderson and Robinson, 2021; Caquard and Griffin, 2019) often cannot be relied on tactile media. Instead, what would it look like if whole-body experience of a tactile map – first encountering the map, reading it, reading it again, setting it down, sharing, storing, retrieving, annotating, referencing, and so on – was taken into account when designing it? Not just as a ‘functional’ tool, but as a part of people’s lives? Participatory methods are helping lend insight into these questions by incorporating the direct insights and experiences of people who are B/VI. The next step in the process, then, is to avoid falling into a trap of complacency by using participatory methods for simple quality assurance and instead advancing/evolving into an era of participatory research.

Two: Creation of tactile maps by people who are B/VI

Tactile map design should be more frequently undertaken by the people who will use the maps. A fundamental characteristic of the digital mapping revolution has been the ability for a far broader swath of society to create, analyze, and share massive amounts of high-quality geospatial data, largely without the need for expensive, specialized equipment or skills. This does not seem to be the case with tactile mapping technology, specifically in terms of tactile map users being able to work with their own data (i.e. sighted users are still largely responsible for creating tactile maps). However, we are beginning to see initial forays into this area. Indeed, tactile map creation is still a specialized skill, although devices such as the augmented reality drawing tool presented by Thevin et al. (2019) show promise for further democratizing the process. But studies such as those by Rassmus-Gröhn et al. (2007), Brittel et al. (2013), Bornschein et al. (2015), and Pandey et al. (2020) – in addition to non-cartographic tactile drawing research, for example, Swaminathan et al. (2017) – all demonstrate that the creation of tactile media on the part of people who are B/VI is not only possible but crucial for the evolution of the discipline. This should not be conceptualized as ‘mere’ drawing but rather in the context of spatial analysis – in the same way that visual mapmaking requires spatial thinking processes to render abstract, often tabular spatial data as geometric forms on a visual document, so too does tactile mapmaking employ spatial cognition, and research needs to attend to that fact.

Three: Continued investigation into low-tech tactile mapping

Research into cutting-edge tactile mapping technology needs to be balanced with sustained attention to low-tech predecessors. Accessible design advocate Liz Jackson will sometimes use the phrase ‘Disability Dongle’ to refer to ‘A well intended elegant, yet useless solution to a problem [people who are disabled] never knew we had’ (Jackson, 2019). This is meant to be a poking fun at overwrought design concepts such as a wheelchair that can climb stairs or a laser-based navigation cane. Tactile mapping research has avoided this trap better than many other fields, but we should not lose sight of the fact that there is still much work to be done with low-tech tactile mapping solutions. Strickfaden and Vildieu (2014), Perdue and Lobben (2016), Vondrakova et al. (2019), and Rahardjo et al. (2019) have all conducted research that does not specifically evaluate the use of any sort of software or digital platform, yet still provide much-needed insights into how tactile maps work. Given that any sort of widely disseminated accessible document (e.g. a tactile map made by the federal government) needs to be usable by as wide an audience as possible, research that avoids reliance on a specific device would help advance that goal.

Four: Increased investigation into the social dynamics of tactile mapping

The social dynamics of tactile mapping are in urgent need of study, for reasons scientific, historical, political, and interpersonal. Any sort of substantive social analysis of tactile mapping is, again, the most glaring absence from this body of literature. Visual cartography has no shortage of sociohistorical research (e.g. Wilson, 2020), but any equivalent in tactile mapping is extremely hard to come by, although not completely absent (e.g. Golledge et al., 2005; Hamraie, 2018; Tanaka, 2019; Weimer, 2017). The simple questions of who makes and uses tactile maps are often taken for granted.

To its credit, research on tactile mapping is geographically cosmopolitan – works cited in this article originate from countries including Greece, Japan, Brazil, Czechia, and many others. However, nearly all these works have been published in anglophone media. Additionally, the global south is generally underrepresented in the literature, and navigation research in particular tends to focus on tactile mapping in urban areas while rural tactile mapping is far less developed. A deeper understanding of the role that tactile maps play in the lives of the people who use them will, on a practical level, lead to the design of more useful maps, or perhaps in some cases propose alternatives, eschewing digital solutions altogether. At the same time, social analysis of tactile mapping should avoid relying on utilitarian justifications for its own existence. There is at this point a long and rich history of tactile mapping to examine and, as has been demonstrated here, a vibrant contemporaneous world that would benefit from such research.

VII Conclusion

In this article, I have provided a review of tactile mapping research between 2002 and 2021. I organized the research according to four prevailing themes: transcription and translation, multimodality, participation, and access and operation. I then outlined a research agenda for tactile mapping research, which was organized according to a further four themes: participatory research, participatory map production, low-tech tactile maps, and social tactile mapping research.

There is also much important research that is directly adjacent to that which I have discussed here but nevertheless will often prove useful to tactile mapping research. Critical and post-representational perspectives on cartography (e.g. Bergmann and Lally 2021; Caquard, 2015) continue to be important for expanding the ontologies and epistemologies of mapping and may prove useful for advancing the notion of representation in tactile graphics more generally. Navigational aids that are not tactile – nor, occasionally, strictly maps – yet are designed for people who are B/VI, constitute a major topical focus for accessible spatial representation. These devices often use audio as their primary modality (e.g. de Borba Campos et al., 2014; Golledge et al., 2004; Oppegaard et al., 2015; Su et al., 2010) but could also use electrical stimulation (Meers and Ward, 2005) or olfactory cues (Jacobs et al., 2015). There is also, of course, extensive research on the geographies of disability and accessibility (e.g. Crooks et al., 2008; Hall and Wilton, 2017; Imrie and Edwards, 2007) that will provide a foundation for future research into the social dynamics of tactile mapping and for developing more robust methodological frameworks.

All of this adjacent research will provide necessary complementary and contrastive material for tactile mapping research moving into the future, which will present yet another set of unique challenges. How will tactile media be affected by pandemic conditions? How does one take advantage of networked tactile mapping technology while mitigating surveillance activity? How can GIScience be conducted using tactile mapping systems? Digital technologies have brought exciting complexity to tactile mapping – let the next 20 years, then, be an era unequivocally characterized by the users of those maps.