Abstract
Introduction
The artistic use of scent and fragrance has a long history in theatrical and, to a lesser extent, operatic productions (see Banes, 2001; Banes & Lepecki, 2007), as well as, on occasion, musical productions (e.g., Macdonald, 1983; Sebag-Montefiore, 2016). There has also been growing interest in the use of scent to augment the displays in museums and art galleries (see Spence, 2020a, for a recent review). In the contemporary era, much of this interest has been spurred on by attempts to enhance the multisensory experience by means of engaging more of the audience’s, or visitors’, senses (Pulh et al., 2008). In the setting of the cinema, there is also a continuing desire to deliver something more (or different) than can be experienced at home (see also B. Barnes, 2014; Hanson, 2013; Lamont, 2016; Malvern, 2018). In fact, back in the 1950s, cinema owners were already worrying about people’s increasing tendency to stay at home in front of the TV, rather than to go out to the movies (Paterson, 2006). As Kaye (2004) notes: “There was a rush of to create technologies to lure customers back to the cinema: 3-D glasses, vibrating seats, and, of course, scented films” (p. 55). 1
However, around the turn of the 19th century, the most pressing question for many cinema owners was how to eliminate the smell of the audience themselves who found themselves in the confined spaces, often with poor ventilation, where early films were typically shown (see J. Barnes, 1983). Once the problem of malodorous audiences had been resolved, the time was ripe for the addition of scent to augment the cinematic experience. And, hard though it may be to believe today, the opinion, at least in some quarters, would seem to have been that adding scent to cinema might well help to elevate the movie watching experience in much the same way that adding color to Black and White films was already starting to do (see Bedi, 2017; Frost, 2006; Hanssen, 2006; Misek, 2010).
2
Note here only the fact that one of the first publicized technologically controlled scented films was shown in the United States in 1940,
3
the same year as the huge global success of early color movies such as
Inspired by the Italian Futurists, Aldous Huxley (1932) also mentioned the olfactorily and haptically enhanced cinema of the future in his dystopian novel
However, from the first attempts to add multiple scents to match what was seen on screen, technological challenges have continued to hamper the use of scent in the cinema (see Gilbert, 2008, for an entertaining review). While there have been several patented attempts to bring scent to cinema, the question of what exactly the goal of this form of sensory augmentation actually was has rarely been considered. The use of scent has often been pleonastic (i.e., redundant with what was shown on the stage or screen; Banes, 2001). That said, the more successful use of scent often references both the action/setting seen on stage or screen but also symbolizes something else as well. For example, J. G. Harris (2007) gives the example of how early audiences watching Shakespeare’s Macbeth in the 1600s would likely have understood the pervasive presence of the sulfurous smell of gunpowder (from the squibs—these were fireworks that would have been set off at the start of the play) in the theater, to connect both to on-stage events and to recent political events in England, namely, the Catholic plot of Guy Fawkes to blow up Parliament—the Gunpowder Plot—that would have been current at the time.
Early History of Malodor in the Cinema
The earliest moving pictures (typically documentaries) were shown on the fairgrounds in the United Kingdom in the 1890s, with one of the main figures responsible being Randall Williams, your author’s great, great grandfather, once known as “The King of Showmen” (see Gashinski, 2011). However, as the projection machinery/technology rapidly developed, it started to become much more cumbersome and hence harder to move from one site to another as the fairground moved around the country from one town to the next. As such, in the opening years of the 20th century, moving pictures started to switch from itinerant attractions that would have been encountered on the fairground to static displays. In fact, during this period, films were increasingly shown in vacant shops or pubs (see J. Barnes, 1983). However, the latter venues typically offered poor ventilation, and low air quality, as compared with what one might have expected to find in the presumably drafty fairground tent. The first purpose-built movie theaters were apparently not much better in terms of their ventilation either. Unsurprisingly, complaints about the smell of the great unwashed masses started to increase at this time. The stench would presumably have been especially noticeable in enclosed spaces, where many people were gathered together in close proximity, such as in the setting of popular early cinema screenings.
What is more, the problem of malodor was apparently much worse in movie theaters than in playhouses/theaters. In part, this was because the former would have been operating for longer each day (c. 7 hours) as compared with just a single show of 2 to 3 hours in the case of the theater—this, at least, according to a report from Richter (1926), a German engineer. Making matters worse, movie audiences may well have gone straight from work to the cinema (i.e., without having time to change). By contrast, those going out to the theater would have been far more likely to get dressed up specially for the occasion. 4 On the basis of his calculations, Richter argued that theatergoers would likely have enjoyed 3 to 4 times more air per person than those in the cinema. It is worth noting that fragrant fountains in the lobbies, and program fans were also, occasionally a feature of theatrical performances in the latter half of the 19th century in London (e.g., see “Fan—Rimmel’s programme fan”, n.d.), again potentially helping to tackle the problem of malodor (see also Reinarz, 2003).
At the 1913 opening of Marmorhaus cinema in Berlin, the fragrance of Marguerite Carré, a perfume by Bourjois, Paris wafted through the building (Berg-Ganschow & Jacobsen, 1987). Given the timing, this is presumably more likely to have been a not so subtle attempt to mask the malodor of the masses rather than an early attempt to give a venue its own signature scent (see Spence, 2020b).
The stench that would arise during movie performances became so unbearable that handheld devices were soon introduced to spray deodorant over the audience’s heads 2 or 3 times per screening. And when this failed to solve the problem, 10-minute “airing breaks” were introduced in a desperate bid to help clear the air (Payer, 2001). Going further, in the 1920s, at Berlin’s Ufa-Palast, an electronically driven flying balloon (a blimp-like device) was launched in the auditorium in order to spray fragrant cologne water during the intermissions. In fact, so widespread was the problem of malodor that one of the early deodorants, going by the name of
Nowadays, during the Covid pandemia lockdown, some people are even apparently craving the contemporary smell of the cinema (Feehan, 2020). Verissimo and Pereira (2013) conducted a study in a Portuguese cinema in Lisbon. A total of 407 moviegoers experienced a cinema complex that either had or had not been scented. The addition of scent was shown to positively impact people’s evaluation of the theater, the environment, and their intention to return when quizzed after watching the movie. Although the article itself is a little unclear, the ambient scent was of cola-lemon (though mint and popcorn aromas were also pretested).
Early History of Synchronized (“Narrative”) Scents in the Cinema
The first “atmospheric” use of scent
5
in a movie theater was by the cinema impresario S. L. (Samuel “Roxy”) Rothafel (1882–1936) in 1908. Rothafel ran the Rialto and Strand movie theaters in New York, as well as the eponymous Roxy, and was also the owner of a cinema in Forest City, Pennsylvania. It has been widely reported that he dipped cotton in rose scent, and held it in front of an electric fan at the latter venue, thereby suffusing the theater with floral fragrance during the newsreel screening of a Pasadena Rose Bowl Game (see Ijsselsteijn, 2003; Paterson, 2006; Ramsaye, 1926, p. 175). Although, as Gilbert (2008, pp. 148–149) notes, there was no Rose Bowl game in 1906, suggesting instead that the scented screening was more likely to have been associated with coverage of the New Year’s Day Rose Parade in Pasadena. In 1929, the manager of Boston’s Fenway Theatre added lilac perfume to the ventilation system designed to coincide with the movie’s title
In contrast to the low-tech scent delivery favored in the theater, cinema screenings gravitated toward a technical (and hence scalable) solution to the programmed delivery of scent instead. On March 4, 1930, John H. Leavell was awarded what appears to be the first patent in the United States pertaining to the automated delivery of scent in the cinema or theater (Leavell, 1930). 6 Leavell’s suggested solution involved cutting notches in one side of the film to trigger a scent (such as the odor of the ocean to accompany an ocean scene) and to cut notches on both sides of the film should a second scent be required. (While this might be taken to suggest that the maximum number of scents that could have be triggered using this approach would have been two, the patent application itself talks about the possibility of delivering multiple odorants.) It is, though, worth stressing that an operator had to manually switch between the odour sources when multiple aromas were to be presented in a film.
The first film to incorporate multiple scents released sequentially to complement the events unfolding on the screen was shown at Paramount’s Rialto Theater on Broadway, New York by Arthur Mayer (1953, pp. 189–190). Gilbert (2008) suggests the unnamed inventor who promised synchronized scent for Mayer’s multisensory screening was none other than Leavell himself. Intriguingly, the scents that were mentioned included rose, honeysuckle, bacon, hospital smell (Lysol), car exhaust odor, and incense (see Wider & Aeppli, 1981).
On January 17, 1939, a patent was awarded to Melvin W. Merz, of Geneva, Illinois for an This invention relates to an aroma diffusing apparatus and its general object is to provide an apparatus which is primarily designed for use in theatres and the like, such as motion picture theatres, for the purpose of releasing odors, to be spread throughout the theatre by air currents from a fan or cooling system in synchronized relation with a scene being shown and to be appropriate thereto or to correspond therewith, such as for example the odor of a perfume is spread when a ball room scene or scene of a lady appears on the screen, the scent or odor of exploding gun powder during a battle scene, the aroma of cooked foods during a restaurant scene, and etc., with the result it will be obvious that my aparatus [
Merz (1939), though, did not seem to realize the challenges associated with distributing scent through a large space. The latter part of his patent application makes it sound as if this would somehow occur instantaneously. He writes: It will be understood from the objects of this invention, that the apparatus which forms the subject matter thereof is arranged whereby the containers are disposed in the path of air currents from an electric fan or the air cooling system of the theatre, so that when the lids of the containers are opened, the odors will be spread throughout the entire area of the theatre, and of course the times of the opening of the lids are synchronized with certain scenes in the picture being shown, for the purpose and in the manner as set forth in the general object of the invention.
A little over a decade later, Emery Imre Stern was awarded the U.S. patent for his smell-delivery system “Electromechanical scent distribution to accompany a motion picture” on August 7, 1951 (Stern, 1951a; see also Stern, 1951b). The application envisages the delivery of “a great variety of scents,” combined with a neutralizing agent, though, once again, it would appear that the system was never used in practice. 8 Following the success of the introduction of “the talkies” (with Al Jolson in The Jazz Singer which came out in 1927 and was the first feature length film with not only music but lip-synchronized singing and speech), the sense of smell would seem to be the next most important in terms of capturing a cinema-goer’s attention. After all, according to Heilig (1955/1992), each sense monopolizes a person’s attention in the following proportions: sight 70%; hearing 20%; smell 5%; touch 4%; and taste 1% (see Figure 1). Notice here how smell is considered by Heilig as being more important even than touch. Heilig goes on to imagine a future for cinema in which: “The air will be filled with odors and up to the point of discretion or aesthetic function we will feel changes of temperature and the texture of things. We will feel physically and mentally transported into a new world” (p. 284).

The order in which the senses focus attention according to Heilig (1955/1992, Figure 7).
In the patent awarded to Laube and Good (1957) on November 19, 1957, entitled
AromaRama
Building on Laube’s patents, in 1959, the “AromaRama” process was released by Charles (“Chuck”) Weiss, with smells piped-in through the air conditioning at the DeMille Theatre, New York.
9
This approach to enhancing the olfactory element of cinema was first used in an Italian travelogue documentary film about the Great Wall of China called
Smell-O-Vision
The next year, January 12, 1960, saw the release of Mike Todd Jr’s version of Jack Cardiff’s thriller
“The olfactory information,” writes Anne Paech, [a Constance-based film historian] “matched by and large the images on the screen, which
However, despite all of the associated expense, this early state-of-the-art version of olfactory cinema was never anything more than a modest commercial success, according to Gilbert (2008) writing almost half a century later. Meanwhile,
It is perhaps worth emphasizing here that Smell-O-Vision did not just have semantically congruent (and hence redundant) visual and olfactory information. Todd and Laube also used several crossmodal tricks to enhance the entertainment value of the scents that were released. So, for example, at one point, a taxi driver was shown drinking what looked like coffee while the audience smelled brandy instead (Sebag-Montefiore, 2015). Meanwhile, in another scene, a person slips at a market and the smell of ripe banana suggests the presumed cause, not shown on screen (slipping on a banana skin being a well-known visual gag). Finally, one of the characters in the movie is associated with a particular scent (namely, the tobacco smoke from actor Peter Lorre’s pipe), and when it is released later in the movie, it acts as a harbinger of the character’s reappearance on screen.
According to Gilbert (2008, p. 169), the organizers of the U.S. exhibit at the 1964 World’s Fair approached Laube about a scented movie project, but eventually dropped it. In subsequent years, several further patents were awarded for the synchronized delivery of odors with motion pictures. For example, in 1974, Westenholz et al. patented their John Malkovich kisses actress Debra Winger on the neck in a run-down hotel in Tangier in the movie “The Sheltering Sky,” and the audience Smells a bewildering blend of oriental perfume, the Sweet Scent of Skin and the basic odor of mould in the hotel room.
Odorama: Scratch and Sniff: John Waters’ Polyester (1981)
In 1981, John Waters introduced
At the same time, however, it is also worth noting that recent virtual reality research has demonstrated how a person’s immersion in a digital experience tends to be enhanced more by the presence of bad smells than by either neutral or pleasant odors (see Baus & Bouchard, 2017). The latter simply just do not seem to do the same job. Indeed, there is now a growing body of research out there to suggest that the brain treats bad smells in an importantly different way from those smells that it finds pleasant or neutral. For instance, we respond significantly more rapidly to bad smells than to good ones (Boesveldt et al., 2010). What is more, we never seem to adapt to bad smells in quite the same way that we do to pleasant or neutral smells, such as the smell of our own home, either (Dalton & Wysocki, 1996; Spence, 2021). It might be argued, then, that Waters intuitively homed in on those scents, namely, unpleasant smells that were likely, in some sense, to be most effective.
Interim Summary
Taken together, early high-tech solutions to scented cinema appeared to have failed because of problems with calibrating the intensity of scent delivery, problems with clearing, or neutralizing, the scent once released, and hence the ensuing problem of a lack of synchronization between what the audience was smelling at a particular moment and the action shown on the big screen (see also Levine & McBurney, 1986; McCartney, 1968; Vlahos, 2006). While the delivery of scent directly to each seat (in the case of Smell-O-Vision), rather than using the air-conditioning (as in the case of AromaRama), undoubtedly helped to address some of these problems, retrofitting cinemas for personalized scent release is likely to have been an expensive process. Low-tech scratch and sniff solutions proved slightly more successful (Waters, 1981), though, that being said, there can be little denying that the appeal of high-tech scent solutions remains amongst those writing in the press (Paterson, 2006). What is more, one of the problems with Odorama-type solutions, where viewers are given a hint to scratch by the numbers appearing on screen referencing which scent to scratch and sniff, risks taking people out of their engagement/immersion in the action, as John Waters himself recognized when talking to Avery Gilbert (2008): “I ask Waters if movie smells can be anything other than a gimmick? ‘You mean for real in a drama? No, I think it will always be a gimmick, because it takes you out of the movie’” (p. 167).
One of the other problems according to director Jack Cardiff when interviewed some decades after the release of
Another problem with the early use of scent in cinema is the sheer number of scents used. Initially, in 1939, Laube had suggested using no more than 10, because more would be “too much for the public’s nose” (see Gilbert, 2008, p. 164), but that number soon increased to 20, then, 30, and beyond. While on the one hand, this can be seen as making the most of what scent-technology had to offer, it also increases the risk of overloading the audience’s nostrils. For instance, 31 odors were sequentially released in the screening of
It may be helpful here to consider the distinction that is drawn by Blankenship (2016) between the “atmospheric” and “narrative” use of scent. The former often involving a single constant scent more commonly used in the theater, at least traditionally. The narrative use of scent became more popular in the cinema setting once technological solutions to facilitate the delivery of multiple different scents started to become available.
A further problem for such technology-led smell delivery systems is that no one has been able to figure out a way of reducing odor perception to some number of odor primitives (as done so successfully in the case of color perception; see also Sebag-Montefiore, 2015). This is something that Heilig (1955/1992, p. 283) clearly failed to appreciate when writing early on that: Odors will be reduced to basic qualities the way color is into primary colors. The intensity of these will be recorded on magnetic tape, which, in turn will control the release from vials into the theatre’s air conditioning system. In time all of the above elements will be recorded, mixed, and projected electronically—a reel of the cinema of the future being a roll of magnetic tape with a separate track for each sense material. With these problems solved it is easy to imagine the cinema of the future.
Recent History of Scent in the Cinema
Despite the early failures, there have been a number of attempts to scent the cinema since the heyday of interest in the latter half of the 20th century (e.g., Anonymous, 2006; Fujiwara, 2006; Sebag-Montefiore, 2015). However, what is especially noticeable is how such ventures typically constitute one-off novelties, seemingly designed more to attract media coverage by promoting the accompanying scent delivery, rather than necessarily a sustained attempt to bring scent to the cinema (in the way that color took over; see Bedi, 2017; Misek, 2010). Many of the latter attempts also reduced substantially the number of scents that were used. For instance,
Patrick Suskind’s (1989) international bestselling novel
Elsewhere, in 2006, one cinema in Tokyo and another in Osaka (Japan) were prepared for an olfactorily enhanced version of Terrence Malick’s 2005 movie
Fujiwara (2006) concludes that: On the whole, the experience was like watching a movie while an aromatherapy clinic was being held in the lobby. Even in my Premium Aroma Seat, I had a hard time distinguishing the scents and often was unsure if a new perfume were being introduced or if a random atmospheric shift had brought a residual scent into stronger focus.
The 2003 feature length cartoon
In 2015,
Edible Cinema: Mixed Reality Solutions
In recent years, there has been growing interest in those cinematic events where the audience gets to experience tasty (and often aromatic) snacks at various points in the film.
17
For instance, Edible Cinema (https://www.ediblecinema.co.uk/) arranged a screening of the movie
Interim Summary
While scent-enhanced cinema is certainly not dead, it is far from a regular occurrence, and primarily seems to be introduced as a means of attracting media attention rather than to really enhance the multisensory viewing experience in the long term. It is typically more of a technical gimmick rather than anything else. What one needs to ask is why they have not been more successful. While the failure of certain early productions (e.g.,
Psychological Problems With Scented Cinema
While commentators have typically focused on
Spence et al. (2017) have also drawn attention to what they have termed the fundamental misattribution error. This is the name given to the fact that even if it could be shown that a viewer’s enjoyment is significantly enhanced by the addition of scent, people are likely to attribute their pleasure to the action they see on-screen, given that we are all visually dominant (Gallace et al., 2012; Heilig,1955/1992; Hutmacher, 2019). Given the cost associated with preparing the relevant scents, not to mention (in some cases) refitting cinemas for scent delivery, the viewing public are unlikely going to be willing to pay for an experience unless they attribute it, rightly or wrongly, to the presence of scent. Given the existence of inattentional anosmia, and the possibly related fundamental misattribution error, it becomes clear how more is demanded of olfaction than was demanded of color, say, when it was first introduced to B&W movies in the 1930s. This may, in part, then help to explain why color movies have been so much more successful than scented cinema. 18
Another relevant psychological factor concerns the fact that while watching a movie tends to be a very social activity, inasmuch as people like to talk about what they have seen, the fact that we all struggle to both imagine and describe smells in words means that such shared discussion is likely to be missing (Arshamian & Larsson, 2014; Majid & Burenhult, 2014; Yeshurun & Sobel, 2010).
The Chemical Sensing of the Audience
According to Heilig (1955/1992): Man is not only a “rational animal” but a “social animal,” as well, and just as he still gets dressed up and goes to a concert hall to hear music he could hear on the radio, he will continue to go to the neighbourhood movies to see the same film he could see at home on TV. (p. 291)
It may not be going too far to suggest that people are sensitive to, albeit unconsciously, volatiles from the other members of the audience, and this, in some small way, contributes to what people appreciate about a shared experience (just as long as they are not too malodorous). 19 According to J. Williams et al. (2016): “the chemical accompaniment generated by the audience has the potential to alter the viewer’s perception of a film” (p. 7). Here, it is worth noting that elsewhere researchers have reported that people are significantly better than chance at classifying whether t-shirts were worn by someone who had watched a funny versus fearful movie (Ackerl et al., 2002; cf. Chen et al., 2006), thus suggesting that it may be possible to isolate the relevant odorants to accompany specific emotions or moods.
Conclusions
Long before anyone had thought about adding positive scents to match what was going on in the movie, the question of how to neutralize the bad smells associated with the presence of so many other customers that was the number one problem. Thus, from the very earliest days of moving pictures, there has been concern/interest around the olfactory contribution to the experience of film (Heilig,1955/1992). However, that interest was initially around how to eliminate the smell of the masses, it has since morphed into the enhancing of multisensory experience by engaging more senses. 20 However, the failure of scented cinema to take off, unlike the phenomenal success of color in movies (e.g., Bedi, 2017; Misek, 2010), is often put down to problems with the technology (see Gilbert, 2008). And indeed technological challenges are especially noticeable in the context of the movie theater (e.g., as compared with the theater; see Banes, 2001). While problems associated with releasing, and then clearing, a sequence of scents in a theatrical setting have been around for more than a century (e.g., Hartmann, 1913; Shepherd-Barr, 1999), the use of scent in theatrical and music settings would seem to have more chance of succeeding both because a smaller number of scents have typically been involved, and also because the low-tech manual administration of scent is much more common (therefore lowering the cost) than the specially wired cinema seats that have, on occasion, been introduced (albeit only into a small number of cinemas). It is worth noting that the incorporation of scent has also been more successful in the context of the theme parks such as the Epcott center (Lukas, 2008), where the films are often specially commissioned specifically for the venue, and hence change much less frequently than the feature films that are typically shown in the cinema.
Returning to the filmmaker, John Waters’ earlier comment that scent in cinema can only ever be gimmicky, Saskia Wilson-Brown, founder of LA’s Institute of Art and Olfaction and a collaborator on the 2015 showings of
It can be argued that the use of scents needs to go beyond the mere redundant presentation of smells associated with that which can be seen (even though this was undoubtedly the focus of a number of the early patents in scent-enabled cinema) if it is to succeed in the long term (see Alberge, 2015). It is not merely enough to smell what one sees (in a redundant multisensory manner), but rather, olfactory cues need to provide something extra, be it a symbolic role (Banes, 2001), jokes (Waters, 1981), and so on. It is also worth noting that bad smells are likely to be more effective than positive scents in terms of their ability to enhance immersion (Baus & Bouchard, 2017). Low-tech scratch and sniff solutions, while easier to implement, risk taking the audience out of the film (Gilbert, 2008). Whereas automatically delivered may not be noticed, due to the phenomenon of inattentional anosmia (Forster & Spence, 2018). And, even if they are consciously perceived by the audience, the latter are likely to attribute their pleasure to what they see on screen, rather than to the presence of the scent(s), this what has been referred to as the fundamental misattribution error (Spence et al., 2017). Furthermore, the fact that people seem unwilling to pay much of a premium for scented cinema means that many of the scents that have been used to date have tended to come across as both cheap and synthetic, further limiting the appeal of this particular form of olfactory augmentation.
Nevertheless, looking to the future, there would appear to be a resurgence of interest in 4D cinema that incorporates scent as part of the enhanced multisensory experience (e.g., B. Barnes, 2014; Hall, 2018). According to one recent press report: “The new 4DX cinemas have been launched after smell was revealed to be the sense British movie fans would most like to have heightened when watching a film” (Hall, 2018). This renaissance supports the suggestion that: “Commercial cinema,” notes the American film historian Patricia Mellenkamp, “bombards the senses of touch, taste and smell, as well as sight and hearing.” Yet this does not necessarily mean that we are moving towards new “sensuous” or even synaesthetic cinematic experiences. (Jütte, 2005, p. 278; Paech, n.d., p. 5)
Furthermore, as more content is viewed at home, what with the rise of services such as Netflix, not to mention the rise in new releases being streamed direct to the home in the era of pandemia, and so on, one might question how/whether olfactory scent delivery will increasingly make it into the home setting. 21 Related to this, virtual reality, gaming, and porn industries have all talked about delivering on olfactory solutions for their respective audiences, be it on home entertainment systems, smart phones, and bespoke hardware (e.g., Cole, 2017; Kerruish, 2019; Natividad, 2016; Tan, 2012; Twilley, 2016; see also Pells, 2015; Ranasinghe et al., 2019; http://cartoonnetwork.com/promotion/smellytelly/; Spence et al., 2017, for a review). At the same time, however, it should be remembered that home-scented cinema was the dream of the Digiscents company. The latter even previewed The Wizard of Oz with scented accompaniment for journalists but note that dedicated scent delivery solution needed for each film given no one has yet figured out what the odor primitives might be (Dusi, 2014; see Spence et al., 2017, for a critical review).
Back in 1999, a journalist from Turin believes that Smell-O-Vision has never taken off because, unlike colour TV, smell has no primaries that can be mixed to make endless combinations. “You cannot create an enormous palate of smells the way you can with [just three primary] colours,” he explains. “And that is a fundamental technological problem.” (quoted in Sebag-Montefiore, 2015)
