At the Institute of Historical Research last night we had Peter Heather on 'Vandal religious policy under Geneseric', which was fascinating even for those of us with almost no knowledge of the Vandals. His main argument was that if you read Victor of Vita’s 'History of the Vandal Persecution' with a sceptical eye, then there is not much actual persecution under Geneseric/Geiseric (439-477), only under his son Huneric. We also got the staggering fact that there were around 600 bishoprics in North Africa at the start of the fifth century (as a result of the Donatist conflict) and the possibility of there being a number of 'slightly bemused Father Ted-like Neo-Platonists' around. In what he admitted was a more speculative vein, Peter thought that Geneseric's religious policy of converting the Vandals to Homoian Christianity (his preferred term instead of Arian) may have been motivated by an urge to find a neutral religion between the amalgam of Vandals (Silings and Hasdings) and Alans that he lead.

Which got me back to thinking again about a problem that often comes up for early medievalists: how much does conversion mean? A lot of contemporary discussions of early medieval conversion suggest the possibility of choosing a religion for political purposes (though Peter thought Geneseric might have been a sincere believer, who attributed his improbable success to his religious choice). Yet does this make conversion too easy? Robert Bartlett has allegedly said that 'changing one's religion is less like changing one's hat and more like changing one's head'. Did conversion mean everything to those who experienced it or nothing?

If you're talking about individuals nowadays converting freely, then it's obviously normally nearer the everything pole: literally a 'turning around'. But what about the early medieval situation of what you might (slightly anachronistically) call 'state-sponsored' conversion? (One of the problems in discussions on such topics is that there is too often a dichotomy of forced v free conversion, as opposed to the reality of a continuum of incentives to change or disincentives to keep your religion. In the same way, Victor of Vita tends to use 'persecution' for any acts negative to Nicene interests). What did a change of religion mean to those who didn't get to make the initial decision?

The problem with assessing this is that in most of the cases we don't get to hear the voices of those opposed to the change. And where we do (as for example in the Christianisation of the Roman empire), it's essentially the elites we hear, who often have an investment in the status of the old religion that is atypical. So I got to thinking about a analogy that's slightly more distant (though not quite as far as modern African conversions): the sixteenth-century English change from Catholicism to Protestantism.

There are some obvious limitations to this analogy. There is better communication and more literacy in the period and the state also has a more effective coercive apparatus. (Though although there are firm religious moves against Catholicism, there is not actual persecution in the sense of widespread danger to life and limb). There's also arguably less doctrinal difference than in a conversion from 'paganism'; on the other hand, there is a substantial change of religious and social practice, in everything from the language of services, and the appearance of churches to the village social events. And we not only have far better evidence 'on the ground' for people's views than in the early Middle Ages, we also have historians such as Eamonn Duffy eager to listen to those opposed to religious change.

Duffy’s explorations (both in The Stripping of the Altars and The Voices of Morebath) suggest that the answer in this case to the effect of the conversion is both everything and nothing. He stresses the religious trauma caused to many all over the country (and particularly in Morebath) and the great attachment felt to the old ways (even to the extent of rebellion). And yet, as he admits, by the 1580s, new traditions were developing around the Elizabethan prayer book. Catholic traditions, in other words, could be largely replaced by Protestant ones in about 50 years.

The English reformation suggests that a state-sponsored change of religion could be traumatic but not traumatising for a culture; short-term confusion, but long-term continuity. In the specific case of the Vandal acceptance of Homoian Christianity, their society may already have been so disrupted by warfare and migration as to make the religious change less significant. As Peter pointed out, it's not clear that most of the Vandals were Homoian by the time they took over North Africa; there simply hadn't been time. The English example also suggests that even widespread resistance to religious change could collapse surprisingly quickly: repression, even without active persecution, looks to erode the base of a religion pretty quickly. Peter referred to the 484 list of bishops 'Notitita provinciarum et civitatum Africae' which refers to 466 bishops of Africa, 88 of whom had 'perished' (which he suspected meant converted from Nicene Christianity). But both the example of English recusants and other cases (such as Christians in the Middle East or diaspora Jews) also imply that even in an atmosphere of considerable hostility and discrimination, small communities of devoted believers can remain in their original faith almost indefinitely, unless they are actively exterminated by the authorities. For those people, too, conversion presumably means everything: for most the impact seems surprisingly transitory.

I’ve been wondering again about the question because I've just read Peter Haidu, The subject of violence: the Song of Roland and the birth of the state (1993), which I thought was going to prove an interesting literary take on the political significance of the chanson de geste. Haidu starts by suggesting the need for interdisciplinary study in a way that seems promising:

The literary scholar or student, asked to read extensively in the kind of writing produced by members of departments of history, economics or sociology, feels unquestionably displaced and imposed upon. In such readings the 'literary' person is dispossessed of the controls and references which provide his or her own mode of power/knowledge. Submitting to the discomfort of this temporary disempowerment is the cost of re-knitting what the institutionalization of our episteme has sundered: the (partial) identity of text with its alterities (p 12).

He goes on:

If the literary scholar and reader must recast the image of the text as ineluctably historical from its very beginning, the historian is required to grasp the complexities that attend contemporary notions of 'textuality'. It is only in so far as the historian goes to school at the modern critic’s table that he or she can begin to make full use of the so-called 'literary' text's historical potential (p 13).

Haidu's work, then, is he claims, nourished by 'extensive reading of medieval history'. though he admits that the historian reader ought to find his views 'somewhat reductive, and at least mildly out-of-date'. The problem is that in practice his views on twelfth century history are in fact not only seriously out-dated but often fairly ludicrous. Essentially his main argument is that the Chanson de Roland shows how the 'subject (monarchy) must dispose of the anti-subject (feudal class) with the least possible overt acknowledgement possible.' (p 175). In other words, we’re back to king v barons and feudal anarchy again.

You can, I suppose, slightly excuse Haidu in that he was writing pre 'Fiefs and Vassals'; on the other hand other works of Susan Reynolds' work were already available (and Haidu even refers to Elizabeth Brown on the tyranny of feudalism as a construct). But perhaps worse is that 'feudalism' doesn't make sense as an answer to what is actually happening in the Chanson. Haidu is correct to say that the Chanson shows the structural weaknesses of knighthood itself as the cause of the disaster of Roncevaux (p 84) But he doesn’t demonstrate that 'feudalism' itself (as opposed to monarchic rule) is seen by the poet as a problem, partly because of the inevitable vagueness about what feudalism is. Haidu seems to be equating it at times with a culture of vengeance, at times with a social theory (which the poem doesn't support) that the king is merely first among equals, at times with a distinction between royal judgement and baronial arbitration.

Is it worth, then, a historian reading a book like this? It doesn't help that the book is prone to jargon and reluctant to explain it; not even my dictionary was much help on 'actantial' or 'designator'. There are some good points lurking, which I think a literary scholar is better at picking up, for example, the way Roland himself seeks to make his dead body into a signifier and the phantom nature of the troops that Charlemagne uses to defeat a large Saracen army to revenge Roland’s death. But my overall feeling is largely that the effort involved in reading the book is barely justified.

The same, of course, could be said about a number of history books (and I'm conscious that some historians, have for example, taken an attitude to the Latin poem Waltharius that would make a literary scholar weep). So why do I feel so frustrated about trying to get to grips with literary scholarship? I think because I find it so tricky to assess the usefulness of specific books easily, which is a combination of several problems.

One is the need to pick up the vocabulary and intellectual framework of not just one discipline but a variety of them. Literary/critical theory isn't just using literary frameworks, but a variety, drawn eclectically from sociology, politics, linguistics, psychology, philosophy and goodness knows what else. So even if you've read quite a lot of literary criticism, it won't help you if you encounter someone following a different critical guru. (Whereas, if, for example, you are a historian interested in using archaeological sources, once you understand the basic techniques and vocabulary, you’'e more or less able to cope with most reports you come across). It doesn't help that some scholars are peculiarly bad at explaining both the theories there using or even telling you their sources for them. It was only when I looked up the 'actantial mode on the internet, for example, that I realised the key role of Greimas' theories in Haidu's book; in the introduction Haidu only mentions him in footnotes, which I had missed. (Similarly, one of the many maddening things about Judith Butler's work is that she devotes an entire chapter to discussing the film 'Paris is Burning' without ever making any attempt to explain the plot to those who haven't seen it).

My second problem is that literary criticism often combines historically useful modes of analysis with historically useless forms. Broadly speaking, a historian is interested either in an author's original intention (however pre-post-modern this may be) or its reception/audience reaction in a particular era. A fair amount of literary theory, however, focuses on how someone today could understand a particular text in the light of modern concerns, which is entirely irrelevant to a historian. In some cases it's easy to separate out modern from possible medieval responses. For example, I regard most psychoanalytical readings as a waste of time: they are only valid for the Middle Ages if you regard the views of Freud, Lacan etc as expressing universally applicable truths, which they quite clearly aren't. (On the other hand I regard materialist readings of literary texts as always potentially historically valid, because it’s hard to think of a period when economic/class interests don't affect literature in some way). But often scholars will combine aspects from several theories together, making it hard to split off the useful from the irrelevant in this way. There are also some theoretical standpoints that can be used in either a historically useful or useless way. For example, there are scholars who apply queer theory to medieval history in useful ways (I'm thinking of people such as Stephen Kruger, Robert Mills and Mark Jordan). But there are also others who either come up with historically implausible examples of queerness or just want to discuss Chaucer's Pardoner yet again.

What this means, I find, is that I have no easy way of knowing at a glance whether a book on medieval literature will be useful to me. The ratio of jargon to intelligible text, can't, because of my particular historical interests in gender, be a sufficient guide. Instead I find myself ploughing through dense verbiage and referring back to my cribs to postmodernism in the hope that somewhere an inspirational idea is lurking. It's this lucky dip aspect to my reading of literary scholarship that I find so frustrating. If anyone can tell me some good shortcuts (whether it is 'never read anything published by Indiana University Press' or 'this is who is worth reading on Occitan literature'), I'd be grateful.

I have been looking again at Georges Duby’s theories about the two medieval models of marriage (in Medieval Marriage: Two Models from Twelfth-Century France (1978) and The Knight, the Lady and the Priest: the Making of Medieval Marriage in Medieval France (1983) and coming to the same conclusions as before: it doesn’t fit the Carolingian evidence. As far as I can tell, no-one has yet demolished Duby’s argument for the twelfth century (though I think there are some holes), so the question becomes: what changed from the ninth century and why?

Duby’s main argument starts from the opposed aims of marriage for laymen (dynastic) and clerics (holy), which results in a wish for different rules on marriage. The three main oppositions are often said to be endogamy v exogamy, control of marriage by families v consent of couple and easy divorce v very limited divorce.

The biggest hole in Duby’s argument is that while it may be possible to argue that there is a ‘church’ viewpoint (given that the period is a time of increasing papal control and the development of systematic canon law), it’s very hard to argue that there is a ‘lay’ viewpoint (even if you restrict it to the high nobility of France). That’s because the wishes of noble laymen are deeply contradictory about what they want from marriage. For example, it’s misleading to say that lay nobles wanted endogamy and that this was in order to preserve property intact. To preserve patrimonies in this way, you really need men to be marrying their nieces or first cousins, and there is very little evidence for such close endogamous marriage in either the early or Central Middle Ages (with the exception of Catalonian counts). In a endogamous society, moreover, you’re expected/encouraged to marry your cousin etc. It’s more accurate to say that what nobles wanted was the possibility, but not the necessity of marrying a distant relative, if it was helpful.

Similarly, parental control of marriage is something that broadly noble parents want, but their noble children (particularly adult sons) don’t want. Heiress hunting and other forms of male hypergamy are far more practicable if you only have to get the consent of the woman and not her family. Even easy divorce, which seems an obvious benefit for laymen, isn’t necessarily so - it’s a problem if your favourable marriage or that of your daughter is broken up by the in-laws. In a sense, what laymen really want is tight rules for everyone but themselves: no divorce but theirs, no consanguinous marriage but theirs, parental consent to all marriages, when they’re the father.

If the clerical model of marriage came to be broadly accepted, at least in theory, by twelfth century noblemen, as Duby rightly claims, it was probably because in practice there were compromises and loopholes in church law which allowed the lay model of marriage to continue. Papal marriage dispensations, for example, provided precisely a way of laymen marrying their distant relatives and divorcing them if needed. Similarly, the church was often prepared to assume consent by the couple unless they kicked up a fuss.

The same tendency to compromise seems to me to be there in the Carolingian period, but in crucially different areas. Firstly, the early medieval church, when it referred to the need for consent to make a marriage, very rarely meant that only the couples’ consent was needed. Almost always, consent by the families was also needed, which is why the crime and sin of raptus included elopement as well as abduction.(There may have been slightly more leeway for widows). In that respect, the Carolingian church stuck firmly to the ‘lay model’.

Secondly, while the limits on consanguineous marriage didn’t change between the ninth and twelfth centuries, the practicalities did. (The argument that there was a change over from ‘Roman’ to ‘Germanic’ ways of counting degrees in the Gregorian reforms doesn’t hold up). Carolingian councils sometimes said no marriage within three degrees, some four degrees, sometimes five, sometimes six or seven, but sometimes as far as relationships were known. But in practice, once you get to four degrees or beyond these probably all boiled down to much the same thing, because in a society almost entirely without written genealogies and relying on oral tradition, four degrees is as much as anyone’s likely to know (and three is more common). (Early medieval royal genealogies/kinglists may seem to be an exception to this, but in fact aren’t really, because they remember only male-line descent and so ignore 90% of relatives.)

In other words, based on oral culture, you might just know who your grandparents’ grandparents were (four generations back), but further back is almost certainly unrealistic, so that the most distant relatives you might know about are your third cousins. (As a parallel I would say, roughly speaking, that except in very close-knit societies, most modern people in the West are close to their siblings (1st degree), know their first cousins (2nd degree) and know of their second cousins (3rd degree). So not marrying your (known) relatives and not marrying anyone within 4+ degrees is much the same in Carolingian society.

What changes is the practice of recording detailed (not just selective) genealogies, which seems to have taken off for nobles only from the eleventh century. (Duby in The Chivalrous Society refers to one example in France from before then). The condemnations of consanguinous marriage seem to have encouraged the drawing up of such charts and, once they’d been recorded in writing, they could be extended into further generations. In other words, it began to be possible for nobles, at least, to know who their fifth degree relatives were, rather than have a convenient ignorance. This change also meant a shift of power over who decided on marriages: control went to whoever drew up the genealogies and used them. In the Carolingian period, in contrast, (and this is made explicit in at least one case), it is the memory of the ‘elders’ that is decisive. In other words, the restrictions gave a chance for any of the parents of the couple (and the elderly relatives more generally) to say: ‘They can’t get married because I remember my granny saying she was related...’ without fear of contradiction by the couple themselves. The rules therefore benefited lay seniores at the expense of iuniores: it’s no wonder that the magnates themselves endorsed it.

The most potential for conflict between different views of marriage in the Carolingian period was on how easy divorce should be. Why did Carolingian laymen (at least in theory) accept the principle of indissolubility? Here, I think you have to say that the ‘church’ did in theory win out. It was harder for Carolingian kings than Merovingian or Capetian kings to get divorced. Unlike in the later periods that Duby sees, there are few examples of kings or nobles blatantly defying church rules on indissolubility. Instead, you see persistent attempts to ‘game’ the system, whether in the infinite excuses of Lothar II or slightly further down the social scale, in men, for example, who claim that they were never properly married in the first place. I can only presume that, as in other cultures where divorce has been almost completely prohibited, enough wiggle room was left for remarriage in practice, while creating a important new holy facade.

So does this mean that you can accept the principle of lay and clerical models for marriage in the ninth century and just say the parameters have changed slightly? I don’t think you can. What you seem to me to have is in many ways an overlapping viewpoint: marriage as monogamous and exogamous and controlled by the family, relatively limited divorce. The differences are at the edges: how much exogamy, when is divorce possible? You don’t get a very different view in the sources written by or for laypeople and those written by or for clerics, as with Duby’s evidence. I’d say it’s less two models than one lay model being pulled in a slightly new direction by some clerics. In which case the real change in the eleventh century is a generation of churchmen refusing to make such compromises, but instead insisting on their own view of marriage (and the power of the church to decide on such matters).

For several weeks, on and off, I have been contemplating a statement by Chris Wickham (in 'The Feudal Revolution', Past and Present 155 (1997)), when discussing contrasts between the Carolingian and post-Carolingian state:

many of us need to find at least some group of powerful people in the past with whom we can feel some sympathy. In my view, in our period [earlier Middle Ages] such a search is fruitless; they are simply not there, and there is no point in regretting their absence.

Leaving aside the specific claim by Chris, this raises the broader question: do you need to have sympathy with the historical figures you study to do so effectively? The example of Chris himself suggests not, and yet you can't help noticing that, for example, most historical work on gays, Jews or women has been done by historians who are gay, Jewish or female. Is there something useful if not necessary to historians in identifications of this sort?

At the most basic level, there is interest. Each historian, regardless of their own particular training and skills, has topics that interest them and topics that leave them cold, and I think it's very hard to go against that. Although many early medievalists are fascinated by land ownership, it does not engage me. Nor does Russian history of any period. I find Chinese history more interesting than Japanese and Italian than Spanish. I'm not sure whether you could make yourself find a period of history or a topic fascinating by willpower, but I'm sceptical. To research history (or even just to learn about it seriously) requires above all a belief that what happened in ninth-century Aachen or the Philippines in the nineteenth-century or wherever else matters, in the face of a world that largely says 'who cares about then?' If deep down you don't really care yourself, it's hard to continue.

But, as Chris shows, you can care deeply about a period without sympathising with most (or possibly any) of the people in it. (I don't know if he feels sympathy with early medieval peasants). So what does sympathising or even identifying with particular historical groups bring? Is there an advantage to being a gay writing about gay history or a black historian of slavery?

I think there are both advantages and disadvantages to that kind of identification. On the one hand, I think it's very difficult to write about actions you can't see the point of at all. I saw a blog post the other day (not on the topic of history) by a vegetarian lesbian who couldn’t imagine being a woman wanting to have sex with a man or wanting to eat a steak. That's honest of her, but it does mean that she should never try to write seriously about straights or carnivores. But everyone has such failures of imagination and in subtly different places. I don't like getting drunk, though I’m not a teetotaller. Although I write in passing about drinking I could not do a good book (or even article) about early medieval drinking culture, because I can't appreciate why someone would want to have drinking bouts, where the fun of it is. Other historians can appreciate such a culture and would write better. Similarly, I can't take polytheism seriously; I can't understand why anyone would believe it, and yet some classicists clearly can.

An important point to make here is that this is about imagination, not necessarily belief. There are plenty of good writers about medieval heretics who aren’t gnostics, and even some who are atheists. It is possible to use analogy to gain this kind of empathy: 'I do not know what it is like to be gay, but I do what it is like to be a marginalised member of society'. I suspect that a really good writer or speaker on a topic can also expand your imagination and sympathies. I don't know whether if I read enough about Greek polytheism, it would come to seem 'sensible' to me or not.

Alongside imagination, there are two other possible areas where identification may be of aid to a historian. One is that it can be easier to spot 'dog whistle' rhetoric or subtexts. If I read a medieval love poem, I read it as a straight, with straight assumptions; a gay scholar may be more likely to pick up the possibility of it talking about same-sex desire. Similarly, I don't know whether a Westerner could have come up with Edward Said’s ideas about Orientalism. For my part, having had a long immersion in Christian thought, I can not only pick up many of the Biblical allusions in a medieval text (which a well-educated atheist could do). I can also spot the Biblical verses and concepts that don't get used, when they could be; notice how few the references are to camels and eyes of the needle, for example.

There are also some experiences which are particularly difficult to imagine if you have not been through them, where analogies do not easily work. I would include combat, for example, the death of a close relative, being a mother. (There are probably several others). Sometimes out of those experiences we have shared with figures in the past, an insight can come which seems to me unlikely to occur otherwise. One of the most interesting cases I've seen was reading Alan Bray, Homosexuality in Renaissance England (1992). One of his key points was the disconnect between Elizabethan England's homophobia and its practice. People saw sodomites as evil outsiders; they did not connect such figures with their neighbours (or even themselves) enjoying themselves with a local youth. I think Bray's revelation of the possible cognitive processes there came from his own experience as a gay man before the legalisation of homosexual relations and I think it would be unlikely to occur to anyone who hadn't had such experiences.

The problem, of course, with identification, is that as well as leading to insights, it can also mislead. John Boswell was accused (a little unfairly) of projecting back twentieth century US gays into the Middle Ages. Sometimes it's easier to get a clear view if you're a little distanced. One of the most insightful historians on 'Germanic lordship' was the Czech Frantisek Graus, and I'd also argue that in some ways it's easier to be a moderate Protestant historian of the Middle Ages than a Catholic one.

But finally, I think there's another issue about sympathy, which is more subtle. A historian often has to contemplate people that he or she can't easily sympathise with, who do horrible things. And yet there is still a need for, if not empathy, or sympathy, at least a quest for understanding. One of my medievalist friends said once that he was interested in why these awful medieval people were so awful. Joanna Bourke has just written a well-received book on the history of rape. To write that must have required more mental fortitude than I have; but it also requires a belief that there is a historical answer to be found. Because one of the biggest enemies of good history is a belief that you already know the eternal answer, for example that rapists rape 'because they can'/'because they are men'. Some of the worst writing on medieval gender (names available on request) is done by those feminist historians who believe that what men do is oppress women. Any change to the social/gender order is tortuously explained as just men finding a new way to oppress women (whereas often, it looks more motivated by struggles between men). Similarly, my heart sinks whenever I start reading historians who claim that all historical behaviour can be explained by the material interests of parties (as economic historians and atheists discussing religion are particularly prone to do). Even Chris Wickham, who, as I've said, is a very good historian, has such blind spots. He writes in the same article: 'Aristocrats were brutal in all periods; it was one of the signs of the aristocracy'. My research looks at the mentalities of noblemen and how they justify their power ideologically. Chris couldn't, I think, do my kind of research, because to his mind it doesn't matter: nobles are eternally just like that. In the end, even if you don't have sympathy with your historical subjects, you have to be willing to listen to them; and be prepared to change your mind if they don't 'say' what you hope they will.

I have just been severely put off an otherwise interesting book on medieval gender (Simon Gaunt, Gender and Genre in Medieval French Literature), by its sloppy use of terminology and lack of historical perspective. He starts quoting Gayle Rubin and Eve Kosofsky Sedgwick on ‘compulsory heterosexuality’ and then comes out with the ridiculous claim: ‘homophobia, like misogyny is an inevitable product of patriarchy’. Just how culture-bound can you get? Ancient Greece and Rome were undeniably patriarchal societies, but you can hardly call them homophobic.

And what on earth does ‘compulsory heterosexuality’ mean anyhow? It may work as a concept for modern Western society, but it certainly does not for many premodern societies. Gayle Rubin (in her classic article ‘The Traffic in Women’), trying to locate ‘obligatory homosexuality’ in prehistory, claims ‘the incest taboo presupposes a prior, less articulate taboo on homosexuality’. But again this theory doesn’t cope with societies (classical or other) that have incest taboos, but no taboos on homosexuality. You don’t have to assume that incest taboos are for eugenic reasons to see that incestuous relations which are potentially fertile are likely to be more socially disruptive than intrinsically non-fertile relations (since they are more likely to be discovered and have more consequences) and therefore that it is those ones which get mentioned. (As an interesting aside, Judaism developed prohibitions on sleeping with one’s grandmother later than with one’s granddaughter).

The term ‘compulsory homosexuality’ seems to me to be unhelpful in several ways. Firstly, what you can often see in the modern world is not ‘compulsory’ but ‘presumed’ heterosexuality. It is assumed that people will be straight and systems/patterns have been developed which follow these assumptions. For example, if I am talking to a woman I’ve just met and she says ‘my partner works at X’, I would probably say: ‘What does he do?’ This is an assumption on my part of the sex of the partner (although one that is likely to be accurate nine times out of ten). If I was wrong, I would expect to apologise and then carry on with the conversation as normal.

Such kinds of heteronormativity may be very irritating for gays, but I don’t think can accurately be called ‘compulsory heterosexuality’. As an analogy, there’s a difference between assuming all people are right-handed and forcing them to be. ‘Compulsory heterosexuality’ seems to me to require more social pressure. But even here, there seems to be a tendency to conflate together several different demands made historically by society on individuals:

1) ‘compulsory procreation’ - that individuals must reproduce
this was common in ancient Mediterranean cultures (e.g. Greek/Roman), but is entirely compatible with any number of homosexual relationships combined with marriage

2) ‘compulsory non-homosexuality’ - that individuals must:
a) not desire their own sex
b) not engage in same-sex sexual acts
these, I would say, were the most common medieval demands. Note that these both allow a positive valuation of asexuality/celibacy.

3) ‘compulsory heterosexuality’- that individuals must:
a) desire the opposite sex
b) engage in opposite-sex sexual acts
these are different from 2) because they mean that asexuality and celibacy are no longer valued. Instead lack of opposite-sex desire/activity is taken as a sign of same-sex desire/activity.

By this definition compulsory heterosexuality is obviously commonplace today, especially in certain macho environments. The interesting question is about the history of this and I think it’s quite complex. There are medieval examples of such attitudes (for example where celibate priests are presumed to be sodomites or when claustration is said to be essential to preserve the chastity of monks/nuns) but I don’t think it’s the dominant discourse until the end of the Middle Ages or even later. I think for men the Reformation did pretty much endorse the unnaturalness of male celibacy and that more or less remained the normal view. For women, I’m less sure. I think asexuality in women retained at least some of its positive value for far longer. I suspect, for example, that it was only in late Victorian period (or maybe even later), that a pair of spinsters living together would come to seem suspicious.

I don’t follow the view of some scholars like James Schultz, who think that the term heterosexuality isn’t useful in medieval studies. I think there is ‘presumed heterosexuality’ inherent in most medieval texts, unlike in classical ones. (Once you can no longer have debates about whether the love of boys or girls is better, you know presumed heterosexuality has come in). But I don’t think ‘compulsory heterosexuality’ is a useful term to be using, unless you think rather more carefully (and historically) than most scholars of gender seem to do about the topic.