Towards a Narratological Approach to Ideology Analysis
The dominant model in contemporary ideology studies is Michael Freeden’s concept-centred “morphological” model. Subsequent scholars like Cass Mudde, have leant their analyses of populism and other far right-adjacent ideological tendencies on this model. This essay is a proposal to adopt an alternative approach to the specific area of ideology study that targets far right ideologies. The proposal is for a more narratological model, stemming from the seminal work of Vladimir Propp, first laid out in his “Morphology of the Folktale”, originally written in 1929, but more or less buried by Stalinist decree, until re-published in English translation in the West in 1958.
Clearly arguing for the possibility of a more productive return on an alternative approach implies two explanations. The first, a critical assessment of the limitations of the current approach and the second, making a case for the potential of the proposed alternative. As an introduction to the stakes involved in the discussion, I first want to look at the brief polemic exchange between Claude Lévi-Strauss and Vladimir Propp that occurred shortly after the appearance of Propp’s “Morphology…” in the West in 1958 (NB the coincidence in the use of the word “morphology” as the self-description for both Propp and Freeden’s contrasting models is the reason for the scare quotes in the first sentence above).
The publication of Propp’s “Morphology of the Folktale” in English in 1958 came at an inconvenient time for Lévi-Strauss. His brand of structuralism was really starting to get traction at this time, and the sudden re-appearance of a key text — one recognised by his mentor Roman Jakobson as a formative structuralist work — was potentially problematic for him. Lévi-Strauss responded in 1960 with an article, ostensibly in the form of a review of Propp’s book “Structure and Form: Reflections on a Work by Vladimir Propp”. The work of this article went far beyond a simple review though. Despite strategic disclaimers on his part, Lévi-Strauss took the opportunity to try and open “clear blue water” between his own brand of structuralism and the unwelcome competition of this newly available alternative from a generation before. Reaching into the Saussurian semiotic terminological toolbox, Lévi-Strauss asserted that Propp’s model was “syntagmatic” whereas his own version was “paradigmatic”. Worse, that Propp’s methodology was indeed “formalistic” as the Stalinists had charged back in the 1930s. He accused Propp of the anti-dialectical sin of separating form from content and ignoring the latter at the expense of the former, along with a host of other alleged theoretical failings. The surprised and offended target of this hatchet job responded with a reply that attempted a rebuttal of the various charges and, in turn, accused Lévi-Strauss of being against any empiricism in method and that he was the one assuming the possible separation of form and content with his “paradigmatic” model which ignored form altogether.
The details of this brief polemic — Lévi-Strauss responded briefly to Propp’s response, playing the wounded party who was shocked by the intemperance of the response to his “innocent article” being treated as the blatant hit-job that it was — an amusing reply which can only really prompt the old saw “methinks the lady doth protest too much…”. Regardless, the merits of this storm in a teacup style spat of professional jealousy between two by now long-dead academics, need not detain us too much longer. The relevant texts and a comprehensive overview by a competent linguist can be found in the 1985 collection of essays by Propp “Theory and History of Folklore” edited by Anatoly Liberman (see the cover above).
In his defence, Propp asserted that his method was an empirical approach that could lend itself to computerised analysis. That may seem a bold assertion for 1958, but in the intervening years since 1928, a good deal of the theoretical work in computational science that we rely on today had been laid down and would presumably have been common currency amongst Propp’s STEM counterparts in the Soviet academic and intelligentsy milieu. To understand that he was basically on target with this gambit, it’s time for a brief sketch of Propp’s model or system.
Propp limited his object of study, not to folktales in general (the book publisher changed the title of the work from the original “Morphology of the Wondertale” in the hope of increasing appeal), but to a specific subset of folktales that he called “Wondertales” — i.e. the stories in a then-canonical collection of Russian fairy stories, roughly speaking. Propp was part of a movement in Russian literary critique that defined themselves against the then-dominant approaches of viewing literature as either a reflection of the cultural-historical conditions of its author’s society, or the psychology of the author themselves. This movement believed that as long as literature was seen as the mirror of another discipline, whether sociology or psychology, then it could not be seen in itself and for itself.
Various means of “breaking the mirror” were attempted. The first iteration concentrated on the use of literary devices in the text. Unsatisfied with the limitations of this first “mechanical” round, led to the next “organic” iteration of looking at how the component devices of a text were articulated into wholes — using the analogy of grammar , specifically syntax — and how characteristics of these wholes define particular genres of literature. Propp’s analytical work on his chosen genre — wondertales — was his contribution to this research project.
The study and classification of folktales was a project already underway outside the confines of what was later called “Russian formalism”. The predominant system was the motif-centred system of the Finnish folklorist Antii Aarne, later revised and expanded by the American Stith Thompson and the German Hans-Jörg Uther, to become the Aarne-Thompson-Uther folktale index, still in use today. To give an example of a motif (a repeated element) in the sense given to it in the index, here’s the entry for the Cinderella family of stories:
510A Cinderella. (Cenerentola, Cendrillon, Aschenputtel.) A young woman is mistreated by her stepmother and stepsisters [S31, L55] and has to live in the ashes as a servant. When the sisters and the stepmother go to a ball (church), they give Cinderella an impossible task (e.g. sorting peas from ashes), which she accomplishes with the help of birds [B450]. She obtains beautiful clothing from a supernatural being [D1050.1, N815] or a tree that grows on the grave of her deceased mother [D815.1, D842.1, E323.2] and goes unknown to the ball. A prince falls in love with her [N711.6, N711.4], but she has to leave the ball early [C761.3]. The same thing happens on the next evening, but on the third evening, she loses one of her shoes [R221, F823.2].
The prince will marry only the woman whom the shoe fits [H36.1]. The stepsisters cut pieces off their feet in order to make them fit into the shoe [K1911.3.3.1], but a bird calls attention to this deceit. Cinderella, who had first been hidden from the prince, tries on the shoe and it fits her. The prince marries her.
Combinations:*This type is usually combined with episodes of one or more other types, esp. 327A, 403, 480, 510B, and also 408, 409, 431, 450, 511, 511A, 707, and 923.
Remarks: Documented by Basile, Pentamerone (I,6) in the 17th century.
Propp’s rejection of this motif-centric approach was that it was too holistic or high level. By including virtually all of the plot as the whole of the item, there was no way of investigating the relations between the individual elements and events of the story. And it made the higher-level categorisations — e.g. “animal stories”, “magic carpet stories”, etc — arbitrary, depending on which elements of the narrative are selected as most salient.
Propp’s more low-level approach was based on analysing the elements of the story — characters, significant objects, events, etc — in terms of their function in the story. If in one story the same function is carried out by a magic animal and in another by a fairy godmother — what matters is the plot-function of the character, not their specific identity or traits. In adopting this functionalist, compositional approach, focussing on the temporal development of the story from beginning to end, Propp’s book proposed a counter-intuitive result, namely that all wondertales had essentially the same structure. One composed of a maximum of 7 characters and a maximum of 31 transitional functions, in strict consecutive order. Not all characters and every one of the 31 functions need appear in every story, but all stories could be fitted to this basic compositional model. Naturally many commentators have since challenged the dogmatically-asserted result of such a rigid invariant structure. But Propp’s methodology and systematisation has had widespread influence.
Problems of Freeden’s concept-centered model
Michael Freeden is a political science academic whose 1996 book “Ideologies and Political Theory: A Conceptual Approach” is rightly recognised as a milestone in establishing ideology studies on a more pragmatic terrain than hitherto found in the field of political philosophy. Freeden’s critique of how political philosophy — including the influential John Rawls’ theory of justice — tends to miss the point of how ideology is used by political movements and parties in the real world of pragmatic manoeuvring and struggling for power and influence. In the old is/ought conundrum, philosophy has a tendency to concentrate on arguing what ideological concepts like justice “ought” to be. Whereas political science has to also look at how it actually “is” in practice, regardless of how inconsistent or “non-ideal” that appears to the philosophical outlook. Freeden’s critique of the political philosophy approach to ideology is openly self-serving or even terrain grabbing, in an inter-disciplinary competition sense. The first chapter that contains the bulk of this preliminary critique is prefixed “Staking Out:…” to make this territorial claim explicit. But in my opinion it is more than fair comment and basically correct.
But that’s not to say that I think there aren’t problems with starting from political science when it comes to militant study of far right ideologies. It’s a case of “if I wanted to go there, I wouldn’t start from here”, basically. It’s not so much that Freeden’s model is somehow wrong in any absolute sense — Freeden makes a point, that I agree with, that one of the “features” of ideology studies frameworks is that they are themselves inescapably ideological in the sense of having to make normative judgements on salient traits and even the nature of power itself. Freeden’s model has to work for the centre ground of political science, what he calls (again correctly, imo) the dominant ideology — liberalism. This is the dominant ideology of governance of our age. For sure conservativism and (social-democratic) socialism are competitors, but liberalism is the ideology they are both in competition with, not so much the other way around. Because political science necessarily has to take liberalism, conservativism and socialism as its centre-ground, Fascism, Communism and Anarchism are consequentially at the margins of the frame. So, more problematically, as Freeden recognises in the final chapters of the book, are important ideological tendencies like feminism and ecologism.
Freeden characterises his model as both concept-centred and morphological (NB his use of morphological shouldn’t be confused with Propp’s earlier use). By concept centred, he means, that liberalism, to use the example most appropriate to his schema, can be defined as an assemblage of concepts, including liberty, progress, equality and rationalism. By morphological, he means that these concepts are not lumped together in an unstructured way, but some concepts are core, some auxiliary and some peripheral. So, for liberalism, liberty, progress, rationality & limited power would be core concepts, pluralism and tolerance could be auxiliaries and laissez faire and welfarism might be examples of peripheral concepts. In his schema there is evolution over time (and in different regions), so peripheral concepts, which tie ideological bodies (parties, movements) to specific strategies and tactics or positions responding to particular moments of historical conjuncture, come and go most frequently. But evolution over time can see previously core concepts moving to more auxiliary or even peripheral position, whereas more distal concepts can find themselves being brought into the core.
Freeden’s model is a flexible model and it gives us a new language to talk about the relative significance and differing roles of core, auxiliary and peripheral concepts, and even which of these have the most practical weight in the ideological evolution of a political collective having to react to the cut and thrust of day to day political struggles. His “Ideologies and Political Theory” is deservedly seen as a foundational work in ideology studies and is pretty much required reading for anyone with an interest in the field. However, I think there are some major problems with it as framework for the specific task of analysing far right ideologies, and the ideological trajectories of groups drifting in that direction.
For Freeden the paradigmatic ideological body is the modern democratic political party, specifically those that have achieved the “maturity” of being parties of power, not just opposition. One of the functions of an ideological system, for him, is that it is able to guide decisions in all areas of policy making a political party in government might have to make, both in terms of executive policies of the various ministries of state, but also the more “soft policy” areas of public communications, media messaging, and generally managing the whole sphere of public political discourse. In other words, the core presupposition of a “full” ideological system, in Freeden’s model, is what Foucault called “governmentality”. There is only one power, state power. The “immature” ideologies of “parties of opposition” or, god forbid, extra-parliamentary political movements, are either peripheral or barely in scope at all. There is no concept of counter-power or the dialectic of constituent and constituted power, as is normal for a political science academic framework, which can only take the perspective of the “ideological superstructure” as the natural vantage point overlooking society. And this is a major dysfunction when studying ideological movements like the insurgent far right that start out as movements of strategic counter-power. At least in the most vital stage of their evolution — from an anti-fascist point of view — the stage *before* they attain power.
The second fundamental limitation of Freeden’s model is its concept-centredness. Freeden quotes Lévi-Strauss frequently in the book, and not uncoincidentally, his framework is very much the latter’s side in the earlier debate between him and Propp. Propp’s criticisms of Lévi-Strauss apply also, to a lesser degree, to Freeden’s morphological model. Whereas Lévi-Strauss’s “paradigmatic” analyses were looking for cultural abstractions that could be reduced to oppositional binaries, it is fair to defend Freeden on the basis that he argues for the essentially contested multiplicity of meanings (echoing Billig, echoing Bakhtin) of his chosen concepts. And the allowance for evolution over time, as touched on above, is certainly far beyond the abstract (ahem) formalism of Lévi-Strauss’s “a-temporal matrix” fantasies.
But even though not explicitly Althusserian in inspiration, Freeden’s concept-centredness is “anti-humanist” to the extent that there are no actors in the model, no human agency internal to the model itself (although of course it is presupposed outside it). Freeden notes, correctly, that ideologies exist as a social force only in their relation to the collective agency of movements or political parties, whatever the roles of individual leaders or theorists. But just because the power of ideology is collective, that does not mean its substance does not incorporate roles and actors, as the means by which individuals are engaged and mobilised into collectives.
In the specific case of far right ideological material, empirically we see a strong role for multiple conspiracy theories. These narrative elements have to presumably be selected, adapted or rejected within some meta-narrative framework that both validates them and itself gains validation from them in return. But to simply bundle this whole area of far right ideology under some catch-all “conspiracism” concept, is side-stepping the task of analytically breaking these narratives down into their components and analysing them “syntagmatically” in their own composition, in their own terms. Unless we break the narratives down by means of a narratological methodology, we cannot advance the analysis of how different narratives relate to each other, in their similarities and differences, and how one may prepare the way for another, more “extreme” (however defined) narrative further down the line. The very process of radicalisation becomes impossible to trace if we restrict our analysis to high-level theoretical abstractions like “pluralism” and, above all, if we do not include the role of actors in our analysis. After all, what is a conspiracy theory but the story of an evil-doing antagonist agency and their wicked deeds? When we study the narratives of far-right adjacent groups or individuals for racist or antisemitic themes, what are we looking for, other than actors in a story and whether the role they play within that story is a heroic one or a malevolent one? When we say that conspiracy theories about George Soros are examples of contemporary antisemitism we are already operating with a model of interpretation that is implicitly actant-centred as much as concept-centred, if not more so.
The final problem with Freeden’s model, from the point of view of research into far right ideology, stems from its “paradigmatic” nature. This relates to Propp’s criticism of Lévi-Strauss that the method of the latter did not lend itself to an empirical approach, particularly one that could be computerised. On the mysterious “black box” nature of Lévi-Strauss’s alleged methodology, I have to defer to Anatoly Liberman’s (1985) dry observation regarding his mystifying paradigmatic analysis of Oedipus:
This is how the “structure” is discovered. According to Levi-Strauss, the first thing to do is to break down each story into the shortest sentences possible and write each sentence on an index card bearing a number corresponding to the development of the story (p. 207). Each gross constituent will contain a relation, but the true constituent units of a myth are not the isolated relations but bundles of such relations. Levi-Strauss’s explanation is far from clear. The first three columns of the “score” consist of eight summarizing sentences. What has happened to all the index cards, of which there must have been several hundred? How were the so-called gross units obtained?
If Lévi-Strauss’s self-described methodology were to be presented in meme form, it would have to be the South Park underpants gnomes meme, with step 1 as index cards and step 3 as paradigmatic analysis, and step 2 being a truly gigantic question mark. There’s no question that Lévi-Strauss’s work has influenced and inspired countless many, but no one has seriously attempted to replicate his “method”, never mind automate it.
Freeden’s approach is of course much more replicable, as befits a teaching academic. But it is still “holistic” or deductive rather than inductive in that the model has to be constructed of the ideology as a whole or a “gestalt”, from a qualitative survey of the entire corpus (or a representative sample) and then individual ideological texts can be compared to the model to how well they fit or not. Propp’s approach is inductive in the sense that individual texts are broken down in terms of relatively objective aspects of their structure, and then the accumulation of break-downs can be analysed to try and draw out general structural patterns. To put it crudely, paradigmatic approaches are a priori and syntagmatic ones are a posteriori, which was what Propp was getting at with his insistence on empiricism. If Freeden’s model gives us a language to talk about the classical universalist political ideologies of modernism, it doesn’t give us a deterministic procedure for either isolating the appropriate abstract concepts or categorising them as core, auxiliary and peripheral.
But rather than getting too much into the philosophical weeds over the inductive/deductive, a priori/a posteriori distinctions between Lévi-Strauss “paradigmatic” and Propp’s “syntagmatic”
competing structuralisms, the advantages of a narratological approach are easiest to grasp at the level of including actors — or “actants” in the jargon — into the analysis. Let’s take time out for a quick example.
The case of QAnon — Trump’s little helpers
The easiest way to illustrate the potential of an actant-centred narratological model is to look at the case of QAnon ideology. In political ideology, as opposed to folk tales, the audience is not just there to be entertained, as mere observers, but is being enjoined to take action in some way. In other words, in political propaganda the audience themselves, rather than being outside the story, is being cast in a particular role within the unfolding narrative. In the more common vein of agitational propaganda, the audience is being cast in the role of the hero of the tale — the valiant defenders of community, nation or race. But if we take Propp’s repetoire of the seven main actor-roles: villain, dispatcher, helper, princess/prize, donor, hero and false hero; we notice something relatively unique about the casting in the QAnon narrative. The role of the hero is already taken by the messianic saviour role attributed to Trump. The QAnon audience are being enjoined to take the role of the hero’s magical helper in the story. A casting that strongly echoes the base evangelical narrative of most of the QAnon audience, where Christ is the hero and the faithful are his magical helpers using the power of prayer to direct the hero’s powers in worthy directions. This crude application of actant-centred narratology, even without doing any real preparatory work, by simply lifting it wholesale from Propp’s domain of folktale analysis and applying it willy-nilly to political ideology (albeit one with the fantastical nature of a darkly morbid wondertale), shows what is potentially possible. The QAnon investigator Travis View, in an interview with the Decoding the Gurus podcast made the observation that unlike any other far right ideological tendency, he felt personally safe to attend QAnon events, where people now recognise who he is, in a way that would be impossible in white supremacist or neo-nazi events. An actant-centred, narratological analysis can give us an explanation of this salient trait of QAnon as a distinct ideology, in a way that a concept-centred account would struggle to do. Similarly, motif-centred or thematic analyses have led people trying to analyse QAnon, noticing the parallels between the adrenochrome narrative and the blood libel of pre-modern and modern antisemitism, to characterise it as a form of crypto-fascism or occult nazism. But such accounts fail to provide an understanding of the non-trivial behavioural differences that View recounts. The bloodthirstiness of the thematic content of QAnon, which is extreme, contrasts dissonantly with the cheerful passivity of an audience faithfully expecting some third party (the army, the “white hats”) to enact their revenge fantasies, rather than have to get their hands dirty doing the slaughtering themselves, in the way the more traditional counterpower models of Klan or fascist ideology would agitate for.
Summing up and next steps
The above segment on QAnon is not a serious narratological analysis. To find the appropriate categories of actants and narrative transition functions we need to do the work of going through iterations of study — analysing conspiratorial narratives to break them down into their elements, then positing general categories by induction and then testing the resulting model against a control group of narratives randomly set aside from the original analysis, in a similar fashion to what we do in training models in machine learning methodologies.
But even before that empirical work, a number of considerations spring immediately to mind. First of all, making best use of existing theoretical resources. Propp’s original morphological model is nearly a century old. However seminal it is, the field of scholarly folkloristics and narratology has hardly stood still in that time. Specifically, in a country like Ireland with a just about still living oral storytelling tradition combined with a developed academia, a review of current work in the field, particularly by people eyeing up the new possibilities afforded by AI and computerised textual analysis, must be worthwhile.
But even in the absence of any worthwhile existing theoretical resources, an obvious question arises in the matter of genre. Recall that one of Propp’s main defences against Lévi-Strauss’s attack, was that his work related only to the specific genre of the wondertale and that other genres, including other types of folktale or mythology — like the Native America myth-tales that Lévi-Strauss was referencing — would have to be analysed in their own right, to uncover the narrative structures peculiar to them. There’s an element of the old confrontation between Cartesian rationalism and scholasticism on the universality of theory in this, or “lumpers vs spliiters” as it is sometimes called, tongue in cheek. This may seem like getting into the weeds, but I think it is necessary to understand one of the major blockers that has historically obstructed leftist engagement with far right ideologies, particularly in the crucial first phase of the 1920s-30s. In this era a number of leftist, specifically Marxist, writers, like Trotsky for example, reacted to Fascist and Nazi ideology by basically throwing up their hands and declaring “these people have no ideology at all”. And if you look at ideology through the lens of a Freeden-style model, it does not look like the total worldview, systemic universalist doctrines that classical liberalism or Marxism typify. But this is a particular way of viewing ideology through the blinkers of an enlightenment modernist viewpoint that inherits from the Cartesian revolution.
One of Descartes major issues with Scholasticism was the Aristotelian teleology that underlay it. For the Scholastics, as each sphere of creation had its own god-given telos, then the theoretical models to study them should be particular to each. Descartes defended a materialist ontology (dodging the charge of atheism via dualism) of a mechanical universe that operated according to laws that, like mathematics, were coherent, consistent and universal. This general paradigm that valid systems are ones that have these mathematical characteristics of being coherent, consistent, and universal has dominated modernist norms ever since, even if the theistic rationalism of Descartes and Spinoza were replaced by agnostic or sceptical empiricism. For the problem in hand, the “lumper” approach would be to gather all far right ideological narratives into a single category and try and find a universal theory that fitted them all equally. The “splitter” approach, is to recognise that far right ideological narratives, follow the more general cultural entertainment narratives from which they inherit, in appearing in a number of heterogenous genres, that each demand particular analytical models corresponding to their particular structures.
To take just two examples we already see in the corpus — the mystery story and the hagiography. The mystery story, or in its most common form, the detective story, is a relatively modern form, appearing only since the introduction of police forces in the early 19th century. Edgar Allen Poe’s “Murders in the Rue Morgue” often being cited as the first of the type. Since its historical recent appearance the mystery plot form has become so predominant in the recreational cultural sphere, that it has not limited itself to its own dedicated genre, but has metastatised across many other genres as well. By contrast, the modern ideological hagiography is a secularised version of the lives of the saints or Vitae sanctorum of the Middle Ages, themselves an adaptation of the hero stories of mythos or folktales. The leftist hagiographies of the late 19th and early 20th centuries, such as the life and martyrdom of Joe Hill and countless other model secular saints of the revolution, have found their distorted and perverted reflection in incel and accelerationist ideological discourses, baptising their terrorist mass-shooters as literal saints and martyrs.
While the older genre of hagiography may be not dissimilar in structure to Propp’s wondertales, the mystery tale has a more complex structure. Again, at the risk of running ahead of the empirical work with idle speculation, for illustration purposes, it’s worth having a brief look at the structural differences.
The crime mystery presents as two tales in one. Not in the sense of the ancient stories within a frame story structure of Arabian Nights, the Oddysey, Canterbury Tales or the Irish Immrama (tales of sea voyages to a succession of other-worldly isles where weird stuff happens — Celtic Star Trek, basically). The mystery lies in the interplay between two plotlines — an inner narrative and an outer one . At the beginning of the outer narrative, the inner story — the story of the crime itself — is hidden. It is the job of the actanct-hero(s) of the outer narrative to uncover the occulted actors of the inner narrative, and sequence of events in the hidden story. Rather than being a simple recursive “story within a story” like a matryushka doll — a series that could potentially be extended indefinitely — the inner story presents itself as an *inversion* of the form of the outer story. So in the inner crime narrative, the perpetrator is the “anti-hero”, who in turn has helpers, auxiliaries, donors, dispatchers and princesses/prizes of their own — all of which must be uncovered by the outer story. The hero of the inner story plays the role of the antagonist of the outer story, and therefore their auxiliaries and helpers appear in that tale as subsidiary villains, often helping to act as decoys in delaying or obscuring the identity of the true antagonist as long as possible for dramatic effect.
Within the exhortation/call to arms format of ideological propaganda, the audience is enjoined to take the role of the hero of the outer tale, the brave detective, intent on “doing their own research” to uncover the hidden tale of crime and unmask the shadowy villain.
Even in advance of the necessary empirical grounding for such speculative accounts, I think its persuasive that much of the mystery genre has had a formative cultural influence on the narratological syntax of conspiracy theories. And these types of narratives have a clearly different structure to the relatively simpler, single-threaded structures of earlier folktales and their derivatives, like the secular ideological hagiography.
There I have to rest the case for an analytical framework that allows for multiple narrative genres, with structures peculiar to each, rather than a grand universal, Lévi-Strauss style “theory of everything”, that by explaining literally everything, ends up explaining nothing.
I’ll wrap up the argument here, in preparation for a more technical proposal for next steps to follow.