• 1 Richardson 2013.

1Many aspects of the way in which we experience, think about, and talk about smells suggest that smells are part of the objective world. First, phenomenologically, smells seem to us to be out there in our environment. The smell of freshly ground coffee reaches our noses, we can sniff around to find where that burnt smell comes from, the pungent odour of gasoline is something that can suddenly hit us and that we can try to stop taking in by pinching our nose.1 On reflection, these smells, scents and stenches seem to be independent of us and our experiences of them: they are entities we can discover, track, get closer to, or try to avoid.

• 2 When talking of the smell of something we might instead mean the individual smell given off by the (...)
• 3 The entities we can smell may be ordinary objects such as cups of coffee and pieces of toasts, the (...)

4Smell Objectivism fits well with this evidence. On Smell Objectivism, olfactory experience presents us with entities in our environment that are objective and do not depend for their existence or nature on individual perceivers, their experiences, or their perceptual systems. For the purposes of this paper, I will focus on smells understood as olfactory qualities or properties, i.e. the properties that account for qualitative similarities and differences among olfactorily perceivable entities. For instance, I will talk of the smell of this cup of coffee as the quality or complex of qualities shared across different cups of coffee having the same smell.2 Accordingly, I will discuss Smell Objectivism as a claim about the nature of smells understood as olfactory qualities. When I lift my cup of coffee and inhale, for instance, I am perceiving a smell – a coffee-y, or more precisely, say, a hazelnut-peachy coffee-y smell – which does not depend on me, my experience of it, or my perceptual system. As formulated, Smell Objectivism is neutral about what sort of entities the bearers of smells are, and whether we perceive those entities in addition to perceiving their smells.3 For all Smell Objectivism says, the smell of coffee may be a property instantiated by the cup, by the stuff in it, or by a cloud of volatile molecules emanating from it.

• 4 Batty’s explicit focus is on the thesis that smells are real and non-relational, but she seems to d (...)
• 5 E.g. Carvalho 2014, Lycan 2014, Young 2016.
• 6 E.g. Richardson 2018.

5In the philosophical literature, Smell Objectivism as the general thesis that smells are objective and independent of perceivers has been defended by Batty (2009).4 More specific versions of Smell Objectivism include views on which olfactory qualities are, or are determined by, chemical properties of stuffs or odour clouds5 and views on which olfactory qualities are objective but irreducibly qualitative, although possibly supervenient on chemical properties.6

• 7 The opponent of Smell Objectivism could of course deny that the initial olfactory qualities of the (...)

10First of all, some cases where we would say that the way the same thing smells changes are not, at a closer look, instances of perceptual variation. In these cases, there is in fact a change in the objective and perceiver-independent properties of the odourant. If you add some basil to a tomato sauce, the sauce comes to smell different: the fresh fragrance of basil now hits your nose. Here the defender of Smell Objectivism has an easy explanation. The sauce smells different because its smell has changed: the objective and perceiver-independent properties present in the two conditions have changed, because a new substance with a smell of its own has been introduced (the basil). Whether olfactory properties are grounded in, supervenient on, or identical to, objective and perceiver-independent properties of odourants, if there is a change in the latter, we can argue that there is a change in the former properties, where these may well be objective and perceiver-independent. This and similar cases are thus not really examples of perceptual variation, and pose no obstacle to Smell Objectivism.7

• 8 For discussion of how similar changes in intensity can be explained in terms of properties of the o (...)
• 9 Some variations in concentration result in a difference in perceived quality, not merely intensity (...)

11Similar considerations arguably apply when there is a change in the concentration or distribution of the odourant or smelly stuff in the environment.8 For example, when we open a window in an attempt to get rid of an unpleasant, stuffy smell, the smell becomes less intense. Letting fresh air in results in a change in the objective and perceiver-independent properties of the source of the smell, of the odour cloud, or of the stuff the cloud or the source are made of. This change can in turn explain the change in the intensity of the smell consistently with Smell Objectivism.9

• 10 Batty 2009: 331-332. Batty focuses on the example of specific anosmia to androstenone, a substance (...)

12Consider now cases that do involve perceptual variation, i.e. where the way something smells changes although the objective and perceiver-independent properties of the odourant, of the stuff the odourant is made of, or of the cloud of odoriferous molecules in the air do not change. Among these cases, some are relatively easy for the objectivist to account for. The first kind of ‘easy’ case is one where the variation in how things smell is explained by a difference in which of the available smells are perceived. One morning, your freshly-made cup of coffee smells intensely coffee-y. The next morning, the coffee has been freshly made just like the day before, but, disappointingly, you cannot smell anything at all; you have a stuffy nose due to a bad cold. The difference in how your coffee smells across the two mornings is explained by a difference in what smells you perceive. As Batty argues, this difference is compatible with the objectivity and perceiver-independence of those smells, because it is simply the difference between perceiving and failing to perceive the coffee-y smell that is there anyway.10

• 11 As a matter of fact, I think that many experiences that are sometimes classified as olfactory illus (...)

14Another class of cases of perceptual variation that are not, on the face of it, especially problematic for Smell Objectivism are illusions and hallucinations. Whatever their causes, these are usually understood as experiences where it perceptually seems to one as if a smell is present even though it is not.11 For instance, someone who suffers from phantosmia may experience a smell of burnt rubber as being present when it is not. There are various ways for the objectivist to account for this apparent awareness of an absent smell, depending on their wider commitments concerning the nature of perceptual experience. For instance, they could appeal to olfactory misrepresentation, mental imagery, or false judgements accompanied by other non-perceptual experiences. Regardless of one’s preference, an account of illusions and hallucinations does not have direct consequences for the nature of olfactory qualities. In the genuine case, the defender of Smell Objectivism would hold that one perceives objective and perceiver-independent smells. In the illusory and hallucinatory cases, they would hold that one fails to perceive whatever objective and perceiver-independent smells might be there, then offer an independent explanation of the fact that it seems to one as if one is perceiving some other smell.

• 12 Lawless 1991, Lawless, Glatter, Hohn 1991.

17What smells one has recently experienced is another factor that can make a difference to how things smell to one. Dihydromyrcenol, for instance, smells more citrusy or more woody depending on whether one has just perceived woody or citrusy smells.12 A more familiar example is that of a piece of 60% dark chocolate, which smells milky and sweet right after one has smelled a piece of darker 90% chocolate, but more cocoa-y after one has smelled a piece of milk chocolate.

• 13 Barwich (2019) argues that olfactory experience generally does not exhibit constancy in the same wa (...)
• 14 E.g. Carvalho 2014, Millar 2020.

19This interpretation of the cases is supported by the fact that perceivers recognise what they are smelling as the same in spite of the changes. While the cake smells different when anchovies are present, it still gives off a recognisably cake-like smell, and one is not inclined to judge that one is now smelling a different cake, or a different kind of thing altogether – a sweet anchovy pizza, say. Even though researchers disagree over the extent and significance of olfactory constancy,13 we are often able to recognise whether or not we are smelling the same sort of thing across changes in the way the thing smells. For instance, when we move closer to a bakery and the air smells more intensely of warm bread, we usually take ourselves to be moving closer to the source of the smell – that is, the one smell of warm bread we are perceiving throughout – rather than taking the smell to be changing.14 That we are often able to make these judgements about the likely constant presence of certain kinds of sources of smells does not mean that our experience does not change across conditions. Instead, the presence of perceptual constancy in spite of perceptual variation suggests that our varying perceptual experiences do not generally mislead us into making false judgements about the sources of smells in our environment. This, in turn, supports the claim that these cases do not involve misperception or illusion.

• 15 Here I present intersubjective cases where subjects differ in olfactory sensitivity, but there are (...)
• 16 Batty 2010: 1154-1155.
• 17 Hawkes, Doty 2017: 30-31.

20Certain intersubjective cases of perceptual variation have a similar structure. Here, properties of the perceiver or their perceptual system affect how things smell to them.15 Perceivers with hyposmia have a lower olfactory sensitivity than the norm, but this condition often does not simply prevent them from perceiving the smells in their environment. Rather, the same thing may smell weaker to them than it does to a normal or normosmic perceiver, and sometimes it may also smell qualitatively different. This is because, as Batty points out, exposure to the same odourant may result in a different pattern of receptor activation in hyposmic subjects, who have diminished olfactory receptor density or receptor sensitivity, and normosmic subjects. And differences in pattern of receptor activation seem to correlate with differences in how things qualitatively smell.16 Subjects with specific anosmias, i.e. anosmias to specific chemical compounds, may also have different olfactory experiences than normosmic perceivers when presented with the same odourants. In most circumstances, a specific anosmia to a certain chemical does not result in a failure to perceive a smell – this is so only when one is presented with that pure chemical on its own. Because the smells we normally encounter are given off by mixtures of many different chemicals, a specific anosmia to one of those chemicals may result in things smelling different to a specifically anosmic and to a normosmic perceiver, although both seem to perceive the smell that is there. For instance, subjects who have a specific anosmia to trimethylamine, a fishy-smelling chemical, tend to eat fish with strong smells that non-anosmic perceivers find revolting.17 This suggests that although both kinds of subjects perceive the smell of fish, fish smells different to them.

• 18 Batty (2009) suggests this interpretation.

21Now, one might wonder whether these intersubjective cases are best understood as involving an illusion.18 One of the two subjects is hyposmic or partially anosmic, which means that their olfactory sensitivity differs from that of normosmic perceivers. However, this is not a good reason for thinking that this subject is having illusions in the cases under discussion, while the normosmic perceiver is not. Being a normosmic perceiver just means having olfactory capacities within a range that is considered normal, where ‘normal’ does not per se have normative implications, and could simply be a conventional categorisation that is helpful for various diagnostic and practical purposes and generally applies to the majority of (human) perceivers.

• 19 E.g. Cohen 2009, Allen 2016.
• 20 Cohen 2009: 31-33.

22Arguments developed in the domain of colour perception tell against treating similar cases as illusions or misperceptions. For instance, among otherwise normally-sighted perceivers there is significant variation in reports concerning hues: certain shades are reported as pure or unique green by some subjects, but as bluish green by others.19 As Cohen (2009) argues, we have no non-arbitrary reasons to take some of the perceivers, but not others, to undergo colour illusions. Criteria such as similarity to the numerical majority, average performance, reference to what is defined as standard for scientific purposes etc. may be useful for certain practical goals, but remain arbitrary as criteria for establishing whether a certain subject is genuinely perceiving the properties in their environment.20

• 21 E.g. Lawless 1997, Keller et al. 2017, Barwich 2019.
• 22 Barwich 2019: 5.
• 23 Hawkes, Doty 2017.

23The same arguments, it seems, apply in the domain of olfactory perception, where differences in sensitivity may be even more common. There seems to be significant variability in the pattern of olfactory receptors each of us is endowed with.21 As a result, our olfactory sensitivity to different odourants may, as a rule, differ to some degree. This might mean that, as Barwich puts it, “individual noses smell the world differently”.22 Both hyposmias and specific anosmias, for instance, are relatively common – for some odourants, they affect up to 50% of otherwise normal perceivers.23

• 24 Burnyeat 1979: 73-74.

26There are certain ways of thinking about perception and the way things appear to us that might lead one to think that objectivist views of sensible properties, such as Smell Objectivism, are threatened by perceptual variation. A very strong assumption that would show the outright incompatibility of Smell Objectivism and perceptual variation is that, if smells are objective, something that has a smell smells the same in all conditions. As Burnyeat argues, something like the assumption that if a certain sensible property is real and objective, then it appears the same in all conditions seems to play an implicit role in various arguments against the reality and objectivity of sensible properties from the pre-socratics to modern philosophers.24 Given this assumption, variations in how things smell would be incompatible with smells being objective and perceiver-independent. But this assumption is obviously implausible. Think just how implausible it sounds in the case of a paradigmatically objective property such as shape: no one would hold that if shape is an objective property, then an object with a certain shape looks the same in all circumstances. Similar considerations apply to a related assumption concerning the way things smell to perceivers. On this assumption, if smells are objective, then something which has a smell smells the same to everyone in a given condition, i.e. everyone who is presented with that thing will have the same sort of olfactory experience. Again, everyone would agree that this assumption is too strong.

• 25 Kalderon 2008, 2011, Hilbert 1987.
• 26 Kalderon 2011: 244-245.

27The reason why these assumptions about the relation between the ways things appear and sensible properties are implausible is that perception, in general, is partial. The partiality of perception is the thesis that what one perceives depends not only on what there is to perceive, but also on one’s perceptual sensitivity and the conditions of perception.25 While the partiality thesis may seem obvious, it has important consequences. As Kalderon argues, if perception is partial, there may be 1) properties of an object that are simply not perceivable; 2) perceivable properties and parts of an object that are “occluded from view” (such as the colour of the back of an object); 3) perceivable properties that are only perceptually available in certain conditions (a colour may not be visible if the light is too dim); 4) perceivable properties that are only available to certain perceivers, given their perceptual sensitivity (colours may not be available to an achromatopic perceiver).26

• 27 Kalderon 2011: 248-249.

28If perception is partial, then not everything there is to perceive about an object is revealed in a single perceptual encounter with it. But then it is to be expected that the object appears different in different conditions and to different perceivers. Different perceptual encounters, in Kalderon’s words, provide different, partial “perspectives” on the perceivable entities around us, making different aspects of what is perceivable available to us.27 Visual perception of three-dimensional objects gives us a straightforward illustration. A cup, for instance, looks different when observed from the top than it does from the bottom. This is because, with a change in point of view, different parts of the cup become available (the handle, the inside of the cup). The cup is genuinely perceived from both points of view, even though not every visible aspect of it is available to one from each point of view.

• 28 There may be conceptions of olfactory objects on which these are three-dimensional individuals, alb (...)
• 29 Kalderon 2011: 247-249, 256.

29How does the partiality of perception apply to olfaction? The example of perception of three-dimensional objects helps only up to a point, as it does not give us a model than can be straightforwardly applied to smells understood as sensible qualities.28 We can, however, turn to the example of colour, which is precisely Kalderon’s target. According to Kalderon, partiality applies to our perception of a property just as to our perception of an object. If perception is partial, there may be visible aspects of a colour that are only available in certain conditions of perception as well as visible aspects of a colour that are only available to certain perceivers in the same conditions of perception.29 For instance, a blue object looks a certain way in dim, orange-y light and another way in midday sunlight. On Kalderon’s view, this is because each lighting condition makes different aspects of the colour available to a potential perceiver. Consider now an intersubjective case of perceptual variation, such as when a blue object looks a certain way to a certain perceiver and a different way to a perceiver with a slightly different visual sensitivity to blue hues. Here different aspects of the blueness are available to different perceivers.

• 30 Ivi: 242-245, 257-258.

30This picture of colour perception as partial has two advantages which are important for our purposes. First, it is compatible with taking colours to be objective and independent of perceivers. The role that perceivers’ properties, such as their sensitivity, play in perception is not that of partly determining the nature of the colours perceived – as on views on which colours are relative to perceivers. Instead, they play the role of determining the perceiver’s partial perspective, and so which aspects of the colours present in their environment are perceptually available to them in certain conditions.30

• 31 Carvalho’s (2014) and Millar’s (2020) discussion of how, in olfactory experience, we can be aware o (...)

35The partiality of perception gives defenders of Smell Objectivism a framework for understanding the varieties of perceptual variation as resulting from the limitations and constraints of our perceptual access to the smells and smelly entities in our environment. And this framework is independently plausible. Barwich (2019) argues, in light of empirical evidence, that variation in the perceivers’ olfactory experiences is grounded in the causal mechanisms of olfactory perception. Our olfactory system is sensitive not only to features of the odourants presented, but also to environmental changes, the perceiver’s own receptor makeup, and contextually salient information. On Barwich’s view, in fact, perceptual variation is a mark of the objectivity of olfactory perception, rather than of its misleading, illusory or purely subjective nature.31

• 32 Kalderon 2008, 2011.
• 33 Broackes 2010: 365-366.

38In the case of visual object perception, the answer is clear: to have different partial perspectives on an object is to occupy different points of view in space, and different perspectives reveal different visible aspects insofar as they bring different parts of the object into view. But as we have already noted, this spatial model will not do for the case of smell. Colour provides, again, a more helpful starting point. One’s partial perspective on a colour seems to be determined by many factors other than one’s point of view, including lighting conditions, the texture and shape of the coloured object, surrounding colours, intervening media, as well as subjective factors, such as one’s visual sensitivity to hue. On Kalderon’s view of the nature of colours, which complements his account of colour perception as partial, colours are ‘multifaceted’, i.e. they have multiple qualitative aspects which can be available in different conditions and to different perceivers.32 An example of a conception of colour that fits this characterisation is offered by Broackes (1992, 2010). On Broackes’ view, the colour of a surface is a complex property that we can characterise by “the way it changes the light falling on it in the process of reflecting it”.33 A colour thus manifests itself differently in different lighting conditions, revealing different qualitative aspects.

• 34 Shoemaker (2006) proposes a new version of this view to address precisely this worry, but it is not (...)
• 35 Kalderon 2011: 257-258.

39Smells too, I propose, are best understood as being multifaceted, qualitatively complex entities. But what does it mean for a smell to have qualitative aspects? One strategy is to think of smell aspects as higher-order perceivable properties of smells. The idea is familiar from the debate on colour perception and has been developed by Shoemaker (2000). On Shoemaker’s view, variation in how a colour looks is explained by the presence of different perceiver-dependent appearance properties of the colour. Analogously, a defender of Smell Objectivism might argue that a smell has multiple higher-level appearance properties, but construed as objective and perceiver-independent. Whenever a smell appears different to one in different conditions, or to different perceivers, different appearance properties of the smell are being perceived. Appealing to perceivable properties distinct from the smells, however, raises various worries. For one, one may worry that smells would turn out to only be perceived in virtue of one’s perception of the related appearance properties.34 For another, one may worry that the appearance properties would be ‘in competition’ with the smell itself for playing the role of what explains the way things smell in perception.35

• 36 Smith focuses on flavours, which are complex properties perception of which involves the contributi (...)

40A more promising strategy is to think of smells themselves as having a complex qualitative structure. Some proposals in the literature on smells and flavours go in this direction. Smith (2008, 2020) argues that the flavour of a wine is a dynamic flavour profile which evolves over time;36 in the same vein, Todd (2018) maintains that smells are variable and acquire new aspects and dimensions; Millar (2020) observes how a perfume may gradually disclose different aspects, from a refreshing top note to a heavy base note. These observations are certainly plausible for many common odourants. But they only apply to cases where the way an odourant smells changes because the odourant itself has changed over time – due to a change in either its chemical properties or the distribution and concentration of volatile molecules it gives off.

• 37 E.g. Kalderon 2008, Broackes 2010, Matthen 2014, Chirimuuta 2017.
• 38 Kalderon 2008, Broackes 2010.

44This conception of smells as qualitatively complex allows us to appreciate an interesting consequence of partiality, which has implications for the epistemology of olfactory perception. As various authors emphasise for the case of colour, perception ordinarily involves a dynamic exploration of the objects and properties one sees.37 If perception is always partial, one could argue that visual exploration in a variety of viewing conditions is required in order to know which colour one is seeing: for instance, one needs to see how a surface ‘behaves’ under different illuminants.38

• 39 Smith 2020: 46.

46Sometimes multiple encounters are required because the smell itself evolves and discloses new aspects. As Smith argues for the case of flavour, “an individual’s flavor perception at a time can only be a snapshot of the flavor of the food or wine we are trying to detect. […] We often have to taste and retaste a dish or a wine to pick out and note all of the flavors it has”.39 But if our perception of a smell is always partial, affected by contextual conditions and constrained by our olfactory sensitivity, a smell, like a flavour, will as a rule reveal different aspects in different encounters – even when the odourant itself does not change.

• 40 Millar 2020.
• 41 Carvalho presents a similar example, but takes it to support the claim that we perceive olfactory o (...)

47Our modes of perceptual engagement with the odourant can make different aspects of the same smell available to us. Our bodily interactions with the odourant, such as the way we inhale and sniff, result in phenomenally different experiences of the same smell.40 Smelling attentively and intently can allow us to discover subtler notes we did not immediately notice (e.g. a hazelnut note in the smell of coffee), which we take to be notes of what we were already smelling.41 Moreover, different ways of attending make a difference to how things smell. We can smell synthetically, by attending to the whole, unified smell of coffee we are presented with; or we can smell analytically, and attend to the hazelnut note, the citrusy note, the cocoa aspect of the same smell. As Smith (2008) argues, both perceptual strategies are equally valid, and expert tasters may be in an epistemically superior position than novices because they can switch between them. On the approach I am recommending, these are examples of perceptual exploration allowing a perceiver to discover multiple aspects of the same smell. Such exploration will lead to even further discoveries if we perceive the same smell in different conditions. For instance, smelling a piece of dark chocolate after smelling a milkier chocolate might make a toasted note of its smell salient; smelling gingerbread spice and garam masala in the same setting might make us notice how they share a sweet, cinnamon-y note.

• 42 Cfr. Smith 2020.
• 43 Of course, this per se does not rule out that some olfactory experiences are illusory. But if they (...)

48Now, one may wonder whether different partial perspectives on a smell, different perceptual experiences of the same smell, really are equally valid.42 I have argued that an odourant may smell significantly different to one in different conditions, or to different perceivers, compatibly with genuine perception of the odourant’s smell.43 Still, we can allow that there is a sense in which some perspectives on a smell are epistemically privileged. For instance, it may be that being in certain conditions of perception – e.g. being in a freshly aerated but not too cold room – and having certain perceptual capacities – e.g. being normosmic rather than hyposmic, and having clear sinuses – are more conducive to judgements about the smells of things that, in our linguistic community, we consider to be correct. In everyday life these judgements will often concern the likely or typical sources of the smell being perceived. For instance, smelling an apple cake right out of the oven is more conducive to judging that the smell one perceives is a smell of apple cake than smelling the same cake when an anchovy pizza is nearby. Certain ways of smelling may be paradigmatic of smells of a certain kind.

• 44 E.g. Richardson 2018.
• 45 I am very grateful to Simon Wimmer for discussing with me many of the ideas explored in this paper, (...)

51I have thus offered the outline of a promising strategy that the defender of Smell Objectivism can pursue in order to account for perceptual variation. Our discussion, however, highlighted some questions in need of further investigation. First, the objectivist needs to develop a full account of smells understood as qualitatively complex entities, clarifying what similarities and differences there are with the case of colour. Moreover, my discussion here remained neutral as regards the precise nature of smells, e.g. whether they are chemical properties of some kind or purely qualitative, irreducible properties. A second question for the objectivist to explore, then, is whether the suggested conception of smells as qualitatively complex fits well within a broadly physicalist view of smells, as Kalderon (2011) argues in the domain of colour, or instead requires that smells are construed as perceiver-independent but irreducibly qualitative.44^,45