Collecting and Connecting Portrait Sittings: A Re-Evaluation of Experiential Feedback in Enhancing Knowledge and Understanding of British Portraiture 1900-1960

by Dawn Kanter

1. Introduction

In this paper, I explain how I used modelling and ontology engineering to address the art-historical subject of the portrait sitting. A portrait sitting is a moment of interaction between artist, sitter (the person who is portrayed) and sometimes others, from which a portrait is typically produced. Textual accounts of sittings, which appear in sources including biographies, diaries and correspondence, give insight into these interactions. They show that participants used portrait sittings to trade on each other’s status, exchange favours and repay debts. To interpret and express this information about portrait sittings, I developed a portrait-sitting ontology – a formal specification of the constitutive elements of sittings, and a portrait-sitting database – information about multiple sittings, expressed using the ontology. The ontology and database are research objects in their own right, which are available as open, citable and interoperable linked data. However, they are also part of a broader enquiry into British portraiture 1900-1960. To understand their role as such, brief introductions to approaches to portraiture and to British portraiture 1900-1960 are necessary.

2. The Case for a Portrait-Sitting Database

Until the 1980s, there was no meaningful difference between the study of portraiture specifically and the study of art generally. Rather, both the history of portraiture and the broader history of art were treated as histories of artists. The problem with this approach is that portraiture is a particularly social practice, as art historians including Marcia Pointon (1993) and Joanna Woodall (1997) have shown. Portraits are representations of people, which tend to use symbols like dress and props to communicate information about the sitter’s position in society. Furthermore, a portrait is usually commissioned. Therefore, its appearance depends not only on the artist’s approach and the sitter’s identity, but on the agenda of the paying patron as well. In the context of this socio-historical approach, attention shifts away from the portrait artwork and towards the portrait contract – the agreement between participants about the content and the purpose of the portrait, the portrait’s intended viewer and so on. 

The portrait sitting is where this contract is negotiated. It is a particular event, in the sense that each sitting has its own set of circumstances and each participant brings with them individual interests and anxieties. But it is also a socially and culturally situated event, as it takes place in a society that has already-established ideas, for example, conceptions of the artistic profession and expectations of gender roles.1 While it is impossible to know for certain what happened at portrait sittings in the past, textual accounts of sittings – found in sources such as biographies, diaries, catalogues and correspondence – give insight into the circumstances of these events, the relationships between participants and the agreements that were made. Effectively, accounts of portrait sittings focus on what sets portraiture apart from other art forms: the extent of its entanglement with social and economic relations. Therefore, they constitute valuable information for the socio-historical study of portraiture.

A context in which portrait-sitting accounts are particularly valuable is the study of twentieth-century British portraiture. It is often argued that portraiture, in the sense of a social and economic practice, did not survive past 1925.2 The reasons given for this ‘death of portraiture’ include that modernist artists and critics were more interested in shape, space and colour than in subject matter; that photography could produce a likeness more quickly and easily than painted portraiture; and that the traditional class of portrait patrons, the hereditary aristocracy, was in decline. To quote the art critic John Berger (1967):

It seems to me unlikely that any important portraits [‘in the sense of portraiture as we now understand it’] will ever be painted again … [In modern portraits by artists such as Cézanne and Van Gogh] … the social role of the sitter is reduced to that of being painted. … [I]t is not [the sitter’s] personality or … role which impresses us but the artist’s vision.

If these arguments are true, it follows that the portrait sitting evolved from a social and commercial transaction to a more one-sided and artist-centric arrangement.3 But was this, in fact, the case? 

I looked to textual accounts of portrait production to test my hypothesis about the sitting. However, a problem with doing so was that information about portrait sittings is generally treated differently to information about portrait objects. Users of cultural heritage resources like museums’ websites and exhibition catalogues have come to expect access to structured data. They can even expect particular data. For example, most art institutions provide for each object in their collections: artist, date, medium, dimensions and acquisition number. 

There are no such expectations or standards for information about portrait sittings. Nevertheless, there are precedents for a portrait-sitting database. The UK Reading Experience Database (UK RED) comprises over 30,000 textual accounts of experiences of reading in Britain and by British subjects between 1450 and 1945; the Reading Europe Advanced Data Investigation Tool (READ-IT) brings together experiences of reading in Europe from the eighteenth century onwards; and the Listening Experience Database (LED) collates over 12,000 individual experiences of listening to music ‘of all kinds, in any historical period and in any culture’. The uses of this information include charting the evolution of reading from a performed, group activity to a solitary and silent one (Colclough 2000); characterising social media ‘stalking’ as a contemporary form of reading (Antonini, Gomez-Mejia and Lupi 2019); and supplementing the perspectives of professional music critics with those of ‘ordinary listeners’ (Brown et al. 2015). 

3. Towards a Portrait-Sitting Ontology

How do the UK RED, READ-IT and LED transform such complex and subjective material as experiential feedback into data, which can be analysed and compared in order to enhance knowledge and understanding of reading and listening as social and cultural practices? Pursuit of an answer led me to the field of knowledge representation, specifically, the practice of ontology engineering. An ontology is defined by Thomas Gruber (1995) as ‘an explicit specification of … the objects, concepts and other entities that are assumed to exist in an area of interest and the relationships that hold among them’. Therefore, a portrait-sitting ontology would be a formal specification of the objects, concepts, entities and relationships which make up knowledge about the sitting. It is an answer to the question: in a set of structured data about portrait sittings (or a portrait-sitting database), which types of information should I include? 

My methodology for constructing a portrait-sitting ontology, which draws on established ontology engineering methodologies including an approach from competency questions (Grüninger and Fox 1995) and ‘Methontology’ (Fernández, Gómez-Pérez and Juristo 1997), is given in table 1.

Table 1: Methodology for constructing a portrait-sitting ontology

Specification phase

1. Motivating scenario: I identified the need for a portrait-sitting ontology.
2. Knowledge acquisition: I collected information about portrait sittings.
3. Competency questions (CQs): I expressed what is important to know about portrait sittings in the form of questions. 
4. Validation of CQs: I sought input from domain experts regarding what is important to know about sittings and adapted my CQs accordingly.

Implementation phase

5. Conceptualisation: I identified the kinds of knowledge (entities, properties and constraints) needed to answer my CQs.
6. Implementation: I formally expressed the entities, properties and constraints identified above, making use of existing ontological resources where possible.4

Evaluation phase

7. Validation: I ensured that computing the data could provide answers to the CQs.
8. Verification: I ensured that a computer could execute queries and make inferences from my data.

4. Specifying Requirements for a Portrait-Sitting Ontology

The specification phase involved establishing the scope of a portrait-sitting ontology.  Ontologies are necessarily selective. For example, the cultural heritage ontology CIDOC-CRM records information that museums need as keepers of cultural heritage objects, such as the condition and legal ownership of objects. Another cultural heritage ontology, VRA Core, is especially useful to cataloguers. It contains terms to describe the relationships between objects, such as part/whole relationships, replicas and reproductions.5 The remit of a portrait-sitting ontology is to fill three existing gaps in knowledge representation:

  • The first gap is for an ontology that focuses on cultural events and practices as opposed to cultural objects. At present there are very few ontologies which do this, of which only one – the Golden Agents ontology, which is used to explore the production and consumption of cultural objects in the Dutch Golden Age – is concerned with visual culture.6
  • The second gap is for an ontology that focuses specifically on portraiture. Although the Golden Agents ontology addresses artistic production, it does not include a model of the sitting as it is not specific to this genre. 

Thirdly, there is a gap for an ontology which not only establishes a body of experiential feedback as valuable information about cultural practices but also considers the ways in which the information has been mediated.7 The provenance of portrait-sitting accounts is an important part of my model, whereas in others – such as the Linked Art Model, which is used by the Smithsonian National Portrait Gallery – the provenance of the data that is modelled is out of scope.8

If the motivating scenario is concerned with ‘why?’, or ‘why construct a portrait-sitting ontology?’, the next stage, knowledge acquisition, addresses the question of ‘what?’, or ‘what knowledge is to be represented?’. I approached this latter question in two ways. Firstly, I searched the literature on portraiture for existing theories about what the portrait sitting is and what is important to know about it. For example, I deduced from Woodall’s and Angela Rosenthal’s work on the portrait sitting that it is a contract between artist, sitter and patron, whereby identity is constructed or reinforced, in exchange for money (Woodall 1997, Rosenthal 1997). Secondly, I examined successive textual accounts of portrait sittings and identified the kinds of information they contain. The 95 accounts of portrait sittings that I collected give insight into, for example, the reasons for producing portraits; the times, dates and locations of sittings; and the relationships between participants.9 These topics – alongside the ideas of contract and exchange, which come from the literature – comprise an indicative list of the constitutive elements of portrait sittings.10

Ultimately, I expressed my requirements for the ontology – in terms of the gaps in knowledge representation that it fills and the particular knowledge that it represents – in the form of questions. Another way of understanding these questions, which are given in table 2, is as enquiries that a set of structured data about portrait sittings should be able to answer. I ensured that these questions were relevant not only to my own research but also to scholarship on portraiture more broadly – in other words, I validated the requirements for the ontology – by conducting a series of interviews with art historians. During the interviews, participants were given short extracts from portrait-sitting accounts to read. They were then asked questions such as ‘In each of the accounts [provided], who would you say is the author of the portrait and why?’. Participants were also asked to reflect on their experiences of portrait-sitting accounts, both at interview and in their own research. The meta-level questions that I sought to answer in this way were: Do you think that these accounts are useful? If so, what do you think is useful about them? What might limit their use, for you?11

Table 2: Requirements for a portrait-sitting ontology in the form of questions, developed in consultation with art historians

1. Contribution – Who contributes to portrait production?

– Whose participation is active and whose is passive?
– Who has influence over the appearance and materiality of the portrait?
– Who can affect the logistics of portrait production?
– Who is responsible for physically making the portrait?
– In all cases, what is the extent of their influence/responsibility?
– Who is the recognised artist of the portrait?
– Who is the recognised sitter?

2. Qualifying criteria – Which events qualify as portrait sittings?

– Which activities were part of the portrait sitting?
– Which activities were part of portrait production more broadly?
– Were artist and sitter in the same space at the same time?
– Was there an agreement or contract?
– Was there correspondence between artist and sitter?

3. Participation – Who participates in portrait production?

– Whose participation is actual and whose is anticipated or imagined?
– What relationships exist between participants?
– Whose participation is optional and whose is essential?
– Whose participation is intentional and whose is incidental?

4. Power relations – Who is in a position of power?

– Who is involved in decision making?
– Who controls what happens to the portrait?
– Who initiates portrait production?
– Who owns a resource that is used in portrait production?12
– Who has privilege because of their class, race, nationality or gender?

5. Psychosocial context – How normative was the interaction?
– Does the interaction disrupt everyday life?
– Do the participants know each other prior?
– What class or cultural differences exist between them?
– In what kind of environment does the interaction take place?
– Is the interaction habitual or novel?
– How important is the interaction?
– What is at stake?
– How do participants feel about the interaction?

6. Credibility – How credible is the portrait-sitting account?

– What is the account’s intended purpose?
– How removed is the account from the sitting?
– How credible is the author more generally?
– What is the extent of agreement with the artwork?

7. Profit and motivation– Who profits from portrait production?

– Who gains money from portrait production?
– Who gains an artwork from portrait production?
– Who gains social or professional status from portrait production?
– Who gets enjoyment from, or is flattered by, portrait production?
– Who gives their time to portrait production?  
– Who expects something from portrait production?
– Whose, and which, expectations are met?

8. Defining the portrait

– What were participants’ intentions for the work?
– How was the artwork received/understood?
– What is the social status of the sitter?
– What was/is the title of the artwork?
– How is the sitter represented?

9. Cross-disciplinarity – How multi-faceted is the portrait sitting?

– Are there overlaps with existing resources e.g. DBpedia?13
– Is the account relevant to the domain of e.g. history?

10. Object properties – What do portraits look like, where are they, and how did they get there?

– What is the format and style of the portrait?
– What is the provenance of the portrait?
– Where has the portrait been exhibited?
– Where has the portrait been reproduced?

5. Implementing and Evaluating the Portrait-Sitting Ontology

In this section, I focus on two of the ten competency questions above, as I discuss the implementation phase of the ontology design process (that is, the identification and expression of the concepts, entities and relationships which make up knowledge about the sitting) and the evaluation phase (the assessment of whether the ontology meets the requirements set out by the competency questions and is technically correct). I discuss these phases together because there is an iterative process going on. Implementation is only truly finished once it is shown, by evaluation, that the ontology is fit for purpose. The purpose of the ontology, for me, is to support more quantitative and systematic analysis of portrait production than is possible at present and, in doing so, to revise and extend knowledge of British portraiture 1900-1960. Therefore, this section also sets out the results of my analysis and the impact of these results on existing theories. Specifically, did portraiture in fact evolve from a collaborative and transactional practice before 1925, to a more experimental and artist-led one after 1925? Furthermore, how do these distinct modes of portrait production – insofar as they exist – map on to our expectations of what ‘traditional’ and ‘modern’ portraits might look like?

5.1 Fundamentals of Ontology Engineering

Since implementation is the formal expression of the ontology, I offer a brief introduction to the core components of ontology design at this stage. The fundamental elements of an ontology are classes, individuals, and properties. A class is a category such as Person or Role.14 An individual is a particular instance of a class, thus, a particular person, for example, John Collier, or a particular role, for example, Recipient in an Exchange of Money.15 A property relates one entity (individual or class) to another. In the statements John Collier is a[n entity of the type] Person and John Collier plays [the role] Recipient in an Exchange of Moneyis a and plays are properties.

The Resource Description Framework (RDF) is a standard model for structuring data that is made up of these kind of relationships between entities, or predicate relations. It is used to describe information about not only the subject matter, in this case, portrait sittings, but also the ontology itself. The RDF Schema (RDFS) and the Web Ontology Language (OWL) are used to express logical axioms, which both contextualise the existing data and allow new data to be inferred. This layered information architecture allows us to construct complex, precise and powerful expressions.

  • Data: John Collier plays [the role] Recipient in an Exchange of Money
  • Metadata: John Collier is a[n entity of the type] Person
  • Axiom: [The class] Person [is] disjoint with [the class] Event
  • Inference: John Collier is not an Event

The use of shared standards such as RDF in ontology engineering supports the discipline’s broader commitment to interoperability, or knowledge sharing and re-use. This commitment plays out additionally in the development of shared ontological resources. For example, upper-level ontologies describe broad concepts which are needed to model knowledge in almost all domains, while ontology design patterns (ODPs) are ‘reusable successful solution[s] to … recurrent modeling problem[s]’ that may be specialised as required. My point is that although the portrait sitting is a new research object – there is no existing ontology or theory of knowledge about its makeup – I did not develop this theory from scratch. Rather, I drew on existing resources including upper-level ontologies (for general concepts), cultural heritage ontologies (for portrait objects), participation and event ODPs (for portrait-sitting events) and bibliographic ontologies (for portrait-sitting accounts), in order to address the multi-faceted phenomenon of the sitting.

5.2 Top-Level Ontological Structure

My portrait-sitting ontology comprises three overarching classes: ObjectRole and Event(see figure 1).16

Object refers to an endurant or persistent entity and Role to a perdurant or temporal one. An example of a persistent entity is the person Augustus John. An example of a temporal entity is Sitter in the Production of a Portrait of Augustus John by William Orpen. The distinction between Object and Role, endurant and perdurant, is important to my ontology because it allows me to make statements about objects (for example, people, artworks and places), which are context-specific. For example, the person Augustus John is a Sitter in the Production of a Portrait of Augustus John by William Orpen, exh.1900 and an Artist in the Production of a Portrait of Dylan Thomas by Augustus John, c.1937-8(see figure 2). In order to leverage the advantages of this ontological structure, I modelled the period 1900-1960 as a series of Events (most importantly, sittings), in which Objects play various Roles. 

Figure 1: My top-level ontological structure

Figure 2: Roles played by Augustus John in two events

5.3 Theme: Contribution

My competency questions on the theme of contribution are addressed through specialisations of the class Role (see figure 3).17Decision Maker is someone who has control over, or determines, one of the factors in portrait production. These factors include the cost of the portrait and the schedule for portrait production (logistical factors); the subjectsetting and size of the portrait (design factors); and the purpose of the portrait.18 A Maker is someone who constructs the portrait, or any aspect of the portrait. A Model is someone who acts as a subject for the Maker. A Reviewer is someone who views and passes judgment upon the portrait-in-progress. Finally, the class Support covers various supporting roles in portrait production, including SupplierMediator and Funder. As these definitions suggest, I assigned each participant a Role – or multiple roles – based on the actions they took during portrait production. This is a different approach to how the roles of Artist and Sitter are assigned, which has to do with who is the recognised author and subject of the portrait.

Figure 3: Roles in portrait production

Using the Portrait-Sitting Ontology allowed me to recognise more contributors to portrait production than is typical, and particularly, more women and working-class participants – especially, but not exclusively, in supporting roles.19 I also found that there was no correlation between the number of participants in portrait production and the date.20 Rather, the number of participants appears to have depended more on the importance of the portrait (whether it was highly anticipated and highly visible) and the habits of the artist and sitter (specifically, how they lived and worked). At some level, these factors are particular to individuals; however, they can also be part of broader trends.

In the case of a portrait of Winston Churchill by Graham Sutherland (figure 4), Churchill’s particularly bad temper and Sutherland’s relative inexperience as a portraitist are particular reasons why the work was reviewed by multiple people during the production process (see table 3).21 More broadly speaking, however, the high number of participants is due to the fact that this was a presentation portrait. By ‘presentation portrait’, I mean a portrait that was commissioned by a committee and funded by multiple fee-payers (or subscribers), who were usually friends or colleagues of the sitter. It was typically then presented to the sitter or an institution on an occasion such as a birthday or anniversary. These kinds of portraits tended to be administratively complicated. They needed somebody to appeal for and to collect subscriptions and someone to liaise between artist, sitter and subscribers. 

Figure 4: Study for a portrait of Winston Churchill by Graham Sutherland © National Portrait Gallery, London

Table 3: Decision Makers, Makers, Models, Reviewers and Supports in the production of Sutherland’s portrait of Churchill

RoleType(s) of role
80th birthday committeeDecision Maker
Adelaide Doughty (wife of Charles Doughty)Support
Alfred Hecht (Sutherland’s friend)Support
Alan Searle (Maugham’s secretary)Support
Charles Doughty (member of 80th birthday committee)Support
Clementine Churchill (Churchill’s wife)Reviewer, Support
Elsbeth Juda (photographer and Sutherland’s friend)Maker
Felix Man (photographer)Maker
Frank McLeavy (chair of 80th birthday committee)Decision Maker
Graham SutherlandDecision Maker, Maker
Hans Juda (Sutherland’s friend)Support
Jane Clark (Kenneth Clark’s wife)Reviewer
Jennie Lee (member of 80th birthday committee)Decision Maker, Support
June Churchill (wife of Randolph Churchill)Reviewer
Kathleen Sutherland (Sutherland’s wife)Support
Kenneth ClarkReviewer, Support
Lord Moran (Churchill’s doctor)Support
Mary Soames (Churchill’s daughter)Support
MPs (donors to birthday presentation fund)Decision Maker, Support
Randolph Churchill (Churchill’s son)Reviewer
Somerset Maugham (previously sat to Sutherland)Support
Wilfrid Evill (Sutherland’s solicitor)Model
Winston Churchill  Decision Maker, Model

Churchill’s dislike of Sutherland’s portrait, which led ultimately to its destruction, is well documented. Sutherland believed that his brief was to portray Churchill as his colleagues were used to seeing him in the House of Commons, whereas Churchill purportedly wanted a grander and more authoritative portrait.22 Because of Sutherland’s attention to the detail of Churchill’s aged body, the portrait has been positioned in a modernist existentialist narrative, alongside works by Francis Bacon and Lucian Freud.23 Its failure is generally attributed to Sutherland’s cruelty or Churchill’s ego; in fact, nobody but artist and sitter tends to be mentioned. An approach from my ontology shows instead that the portrait is of a type (the presentation portrait) which existed since the mid-nineteenth century and which was particularly collaborative in the sense that it involved multiple commissioning agents and multiple funders.24 As befits this collaborative mode of production, I contend that responsibility for the portrait’s failure lies with the Mediators (Supports) on the 80th birthday committee, for example, Jennie Lee and Charles Doughty, who failed to communicate an agreed brief.

5.4 Theme: Profit and Motivation

The second theme that I explore here is profit and motivation, specifically, the concept of exchange. Using the classes GiverRecipient and Commodity and the properties gives and receivesI was able to describe what participants contributed to portrait production and what they received in return. By doing so, I found that participants in portrait sittings exchanged a great variety of commodities. Not only were portraits made in exchange for money, but participants also traded in books, cigarettes and goodwill gestures, for example, hospitality and recommendations. In fact, these different kinds of exchanges often took place together. For example, a portrait of the scientist Michael Foster by John Collier (figure 5), was produced in recognition of the sitter’s friendship with the artist’s father-in-law, in other words, it was a gesture of goodwill (The Lancet, 1910).25 Nevertheless, the artist also made money from the portrait, as he was paid to produce a copy for the Royal Society.26 Thus, the portrait functioned in both a market economy and what we might call an ‘economy of regard’ (Offer 1997) or a gift-giving economy.27 

Figure 5: Portrait of Michael Foster by John Collier © National Portrait Gallery, London

Another portrait – a bust of George Bernard Shaw by Jacob Epstein (figure 6) – was also made to recognise a friendship and also in exchange for money (Epstein 1955). In this portrait, the sculptor’s marks are very visible, whereas Collier’s portrait is more highly finished. And where Collier shows Foster in his role as scientist and lecturer, Epstein seems to focus more on Shaw’s particular features and expression. Thus, modelling portrait production exchanges allowed me to identify similarities between portraits that are visually and materially different.

Figure 6: Portrait of George Bernard Shaw by Jacob Epstein Photograph © National Portrait Gallery, London

There is one type of exchange that I encountered surprisingly often: the exchange of a portrait for another self-made, often autographed, work.28 One such exchange was between the writer G.K. Chesterton and the medallist Theodore Spicer-Simson. At the end of Chesterton’s sittings to Spicer-Simson, the writer gave the artist a signed copy of his book, in return for a portrait medallion (Spicer-Simson 1962). One interpretation of such exchanges is as attempts by sitters to construct themselves as artists’ professional equals. This interpretation is supported by the fact that the exchanges were often accompanied by conversations in which artists and sitters traded expertise and compared practices.29 Another interpretation is that participants treated the objects in question as souvenirs or tokens of affection, since they represent a person in various ways. A signed work or a cigarette is an index trace of a person, while a portrait is not only an index but also an icon. This interpretation is reinforced by inscriptions such as the one in Chesterton’s book: ‘To T. Spicer-Simson, who modelled me’. 

The presentation portrait type combines these notions of expertise and affection. It affirms professional identity in the sense that it typically recognises the sitter’s lifetime achievements.30 However, presentation portraits also reinforce more personal connections. They were typically commissioned by admirers of the sitter, who sometimes gifted their signatures alongside portraits.31 Moreover, in exchange for their financial contributions, subscribers could receive a portrait reproduction.32 The presentation portrait thus not only challenges understandings of the artist as sole author; it also epitomises an understanding of portraiture in terms of interactions and exchanges, which are inextricable from broader social and economic structures (namely, market economies and economies of regard). Moreover, as I alluded to in my discussion of Sutherland’s Churchill, further analysis of collections and exhibition data shows that the presentation portrait type emerged around the 1840s and endured until the 1940s (see figure 7).33 Therefore, drawing attention to this type is one of several ways in which my method led me to propose an alternative periodisation of portraiture, which challenges the notion of the genre’s death by 1925.

Figure 7: The chronology of presentation portraits

6. Conclusion

 To support a socio-historical study of twentieth-century British portraiture, I developed a portrait-sitting ontology – a formal specification of the constitutive elements of sittings, and a portrait-sitting database – information about 65 sittings, expressed using the ontology. I used the Portrait-Sitting Ontology and Database to reinterpret particular portraits in terms of collaborations and exchanges and, as a consequence of doing so, to identify similarities between portraits that are visually and materially different. As such, my digital methodology supports new groupings of portraits and new periodisations of portraiture. It shifts attention away from stylistic differences between portraits and towards shared social and cultural ideas (for example, understandings of identity, celebrity and gift-giving) that underpin particular types of portrait practice and function. 

7. References

‘Presentation to Mr Holman Hunt’ (1900) Times, 5 November, p 10.

‘Presentation to Sir Jesse Boot’ (1910) Derby Daily Telegraph, 27 January, p 2. <>

Antonini, A., Gomez Mejia, G. and Lupi, L. (2019) ‘All We Do is “Stalking”: Studying New Forms of Reading in Social Networks’, Proceedings of the 30th ACM Conference on hypertext and Social Media, pp 111-115.

Berger, J. (1967) ‘No More Portraits’ in T. Overton (ed.) Landscapes: John Berger on Art. London and New York: Verso, 2016.

Berthoud, R. (1982) Graham Sutherland: A Biography. London: Faber and Faber Limited.

Blackett-Ord, C. (n.d.) Sir Edward Sharpey-Schafer.

Brown, S., Barlow, H., Adamou, A. and d’Aquin, M. (2015) ‘The Listening Experience Database Project: Collating the Responses of the “Ordinary Listener” to Prompt New Insights into Musical Experience’, International Journal of the Humanities: Annual Review, vol.13, pp 17-32.

Cayzer, E. (1999) Changing Perceptions: Milestones in Twentieth-Century British Portraiture. Brighton and Portland: The Alpha Press.

Cohen, P.M. (2008) ‘On the Death of Portraiture’, Yale Review, vol.91, no.3, pp 62-72.>

Colclough, S. (2000) ‘Recording the Revolution: An Introduction to the Reading Experience Database’, Siegener Periodicum zur Internationalen Empirischen Literaturwissenschaft, vol.19, no.1, pp 36-55.

Eaker, A. (2018) ‘The Scene of the Sitting in Early Modern England’, Association for Art History, vol.41, no.4, pp 650-679.

Epstein, J. (1955) Epstein: An Autobiography. London: Hulton Press.

Fernández, M., Gómez-Pérez, A. and Juristo, N. (1997) METHONTOLOGY: From Ontological Art Towards Ontological Engineering’, AAAI Technical Report SS-97-06.

Gruber, T. (1995) ‘Toward principles for the design of ontologies used for knowledge sharing’, International journal of human-computer studies, vol.43, no.5-6, pp 907-928. <>

Grüninger, M. and Fox, M.S. (1995) ‘Methodology for the Design and Evaluation of Ontologies’, Proc. IJCAI’95, Workshop on Basic Ontological Issues in Knowledge Sharing.

Holmes, C. (1936) Self and Partners (Mostly Self): being the Reminiscences of C J Holmes. Constable.

Jones, J. (2008) ‘Inconvenient truths’, The Guardian, 28 August. <>

Juda, E. and Courtney, C. (2001-3) Elsbeth Juda: On Graham Sutherland’s battle to paint Winston Churchill’s portrait.

Kanter, D. (forthcoming) Collecting and connecting portrait-sittings: a re-evaluation of portrait-sitting accounts in enhancing knowledge and understanding of British portraiture 1900-1960, PhD thesis, The Open University.

Liu, D., Bikakis, A. and Vlachidis, A. (2017) ‘Evaluation of Semantic Web Ontologies for Modelling Art Collections’, New Trends in Databases and Information Systems: ADBIS 2017 Short Papers and Workshops, AMSD, BigNovelTI, DAS, SW4CH, DC, Nicosia, Cyprus, September 24–27, 2017, Proceedings 21, pp 343-352.

Lynton, N. (2000) ‘Portraits from a Pluralist Century’ in Gibson, R. and Lynton, N. Painting the Century: 101 Portrait Masterpieces 1900-2000. London: National Portrait Gallery. 

Offer, A. (1997) ‘Between the gift and the market: the economy of regard’ The Economic history review, vol.50, no.3, pp 450-476. <>

Pointon, M. (1993) Hanging the Head: Portraiture and Social Formation in Eighteenth-Century England. New Haven and London: Yale University Press.

Rosenthal, A. (1997) ‘She’s got the look! Eighteenth-century female portrait painters and the psychology of a potentially ‘dangerous employment’’ in J. Woodall (ed.) Portraiture: Facing the subject. Manchester: Manchester University Press, pp 147-166.

Schama, S. (2015) The face of Britain: a history of the nation through its portraits. Viking.

Simon, R. (1987) The portrait in Britain and America: with a biographical dictionary of portrait painters, 1680-1914. Oxford: Phaidon.

Spalding, F. (1991) ‘The Modern Face 1918-1960’ in R. Strong et al. The British Portrait, 1660-1960. Woodbridge, Suffolk: Antique Collectors’ Club, pp 387-421.

Van Wissen, L., Zamborlini, V., Van Den Heuvel, C. (2021) ‘Toward an ontology for archival resources. Modelling persons, objects and places in the Golden Agents research infrastructure’ Digital History Berlin, 11 January [Blog]. <>

Whitworth, G. and Gomme, A. (1932) ‘Mr. William Poel’, Times, 22 July, p 10. <>

Woodall, J. (1997) ‘Introduction: facing the subject’ in J. Woodall (ed.) Portraiture: Facing the subject. Manchester: Manchester University Press, pp 1-25.

About the Author

Dawn Kanter is a PhD student at The Open University. She previously studied architecture and art history at the University of Cambridge and The Open University respectively, and worked in digital roles at Imperial War Museums and the National Portrait Gallery. In her doctoral research, Dawn draws on both her academic and her professional background by developing a portrait-sitting database, which she uses to give new insight into twentieth- century British portraiture. The project, entitled Collecting and connecting portrait sittings: a re-evaluation of portrait-sitting accounts in enhancing knowledge and understanding of British portraiture 1900-1960, is funded by the Open Oxford-Cambridge AHRC DTP.

  1. See, for example, Eaker 2018 for the artistic profession and Rosenthal 1997 for gender roles.
  2. The ‘death of portraiture’ is typically treated by authors who reject this notion less as an argument put forward by a particular scholar or group of scholars than as a trope that brings together various sentiments, including fear for the future of naturalistic painting; declining respect for professional portraitists as artists; and nostalgia for the upper-class lifestyles of which portraiture was part. See, for example, Cayzer 1999, Lynton 2000. Some key contributing factors to the collective notion of the death of portraiture can nevertheless be identified.
  3. Frances Spalding (1991) and Elizabeth Cayzer (1999) make this argument explicitly. According to Spalding, modern portraiture was no longer so much a ‘bargain’ between parties as a ‘creative meeting’, the product of which might not ‘[follow] custom or accepted style’. Similarly, in Cayzer’s view, before 1914, patrons requiring a portrait could approach one of several ‘well-established ‘names’ … [who] understood the requirements and pit-falls of portrait commission’. This statement implies that after 1914, it was less clear who the major portraitists were and what one might expect by engaging such a person.
  4.  The formal expression may be called ‘formalisation’, while the use of existing ontologies and ODPs is ‘integration’.
  5. An evaluation of CIDOC-CRM, VRA Core and the Europeana Data Model, which informs this comparison, was carried out by Danfeng Liu, Antonis Bikakis and Andreas Vlachidis (2017). CIDOC-CRM stands for ‘CIDOC Conceptual Reference Model’. VRA Core stands for ‘Visual Resources Association Core categories’.
  6. Examples of event-centric ontologies outside the domain of visual culture are those used by the UK RED, READ-IT and LED. The Golden Agents ontology itself is available via GitHub. My use of this ontology is limited for various reasons, including its having been developed as recently as 2021.CIDOC-CRM is another event-centric model in the sense that every property of an object (for example, its condition state) is treated as specific to the circumstances in which that property was recorded (for example, the date that the object’s condition was assessed and the identity of the assessor). However, typically, the focus of projects that use CIDOC-CRM remains the cultural heritage object rather than the events, which involve cultural heritage objects, themselves.
  7. In the manner of ontologies such as the Historical Context Ontology (HiCO) and the Reconstructions and Observations in Archival Resources ontology (ROAR). This gap is addressed to some extent by the Golden Agents project (Van Wissen, Zamborlini and Van Den Heuvel 2021).
  8. The Smithsonian National Portrait Gallery is part of the American Art Collaborative.
  9. Unpublished accounts of sittings (for example, those contained within private correspondence) do not appear on the Portrait-Sitting Database website.
  10. In this respect, knowledge acquisition is closely linked to conceptualisation.
  11.  Interview questions and transcripts
  12.  This may be a physical resource, for example, a canvas, or a non-physical resource, for example, expertise in painting.
  13. DBpedia is an established data authority.
  14. The semantics of natural language are not as precise as the semantics of ontology engineering. I have used italics to indicate that I am referring to a precise term in the ontology, as opposed to a concept more broadly. The names of classes and individuals are also capitalised, whereas the names of properties are not.
  15. More precisely, a class represents a set (in the sense of mathematical set theory) of individuals.
    1. As such, it takes inspiration from the ParticipantRole ODP

    In the diagram, each box represents a class in the ontology. The upwards arrow indicates a sub-class relationship. For example, pse:Object is a sub-class of dol:endurant. The prefix ‘dol’ refers to the Descriptive Ontology for Linguistic and Cognitive Engineering (DOLCE), an upper-level ontology. The prefix ‘owl’ refers to the Web Ontology Language. The prefix ‘pse’ refers to the Portrait-Sitting Ontology.

  16. The diagram is selective as opposed to comprehensive.
  17. Factor is itself a Role. It is defined as ‘An aspect of portrait production [more specifically, a Role in portrait production] about which decisions are made’.
  18.  I found that the mean number of participants (specifically, unique Actors) per sitting is not between two and three, as we might expect (two implying artist and sitter and three implying artist, sitter and patron) but rather, four, with the number of participants per sitting ranging from two to 24. See supporting data (Participants per sitting). Across my supporting data, information from unpublished portrait-sitting accounts (or the records which include those accounts) is redacted.
  19. Correlation coefficient = 0.23 to 2 decimal places. See supporting data (Participants by date).
  20.  Information about the portrait’s production may be found in Roger Berthoud’s biography of Sutherland (1982) and an interview with one of the participants in this event, Elsbeth Juda (2001-3).
  21.  See, for example, Berthoud 1982 and Schama 2015.
  22. According to the art critic Jonathan Jones (2008), all three artists tell ‘inconvenient truths’ about their sitters. Cayzer (1999) describes their work in terms of ‘scrutiny’ (Bacon); ‘naked[ness]’, meaning without ‘airs and graces’ (Freud); and ‘probing’ (Sutherland).
  23.  I return to the chronology of the presentation portrait in the following section.
  24. Press cutting in NPG archive, Registered Packet 1869.
  25. The Royal Society website notes both that the portrait was ‘presented by subscribers’ and that there existed a ‘Michael Foster portrait fund’ (to which, it is inferred, subscribers contributed).
  26. ‘Exchange is not only an economic transaction, it is also a good in itself … usually in the form of a personal relationship. … Gift exchange has two elements: the gains from trade and the satisfactions of regard’. Offer 1997 (emphasis my own).
  27. See supporting data (Personal Objects).
  28. See, for example, accounts of Thomas Hardy’s sitting to Spicer-Simson for a portrait of 1921 (Spicer-Simson 1962) and Joseph Conrad’s sitting to Epstein for a portrait of 1924 (Epstein 1955).
  29. See, for example, accounts of the production and presentation of a portrait of William Holman Hunt by William Blake Richmond (‘Presentation to Mr Holman Hunt’ 1900) and the production and presentation of a portrait of William Poel by Henry Tonks (Whitworth and Gomme 1932).
  30. See, for example, accounts of the production and presentation of a portrait of William Holman Hunt by William Blake Richmond (‘Presentation to Mr Holman Hunt’ 1900) and the production and presentation of a  (‘Presentation to Sir Jesse Boot’ 1910). 
  31. See, for example, accounts of the production and presentation of a portrait of Harold Lee-Dillon by Sydney Carline (Holmes 1936) and the production and presentation of a portrait of Edward Sharpey-Schafer by Charles d’Orville Pilkington Jackson (Blackett-Ord n.d.).
  32. See supporting data (Presentation portraits).