With the general resurgence of interest in feminist issues which has occured over the last couple of years, it seems timely to take a closer look at the current state of gender in archaeological theory. Within this lies a wish that we use this momentum of current engagement with gender (in)equality to examine not only our internal practices, but crucially, how we construe and present gender in the past.

This is the heart of the argument presented here: that we need to readdress this point, to ensure we are not miscommunicating a false belief in the fixity of modern gender roles. Whilst this is certainly not a new objective, having in fact formed the core of gender archaeology since its earliest inception (Conkey and Spector 1984; Conkey and Gero 1991; Gero and Conkey 1991; Spector 1991; Wylie 1991a, b; Gilchrist 1999), it remains an objective which has not yet been fully achieved. It therefore requires urgent and renewed attention, not only in the light of the knowledge that the stories we tell about the past influence the present but also with a focus on why we are still in a situation where this is a point for discussion considering gender archaeology has been present in the published archaeological debate since the early 1980s.

This article will make three central points regarding gender and archaeology, with the hope of inspiring introspection and change in the way in which our discipline regards and deals with questions of gendered identities and roles. The first of these is a genuine concern that despite the excellent work done within gender archaeology in the last four decades, questions of gender still remain on the margins of many archaeological enquiries. Indeed, the very fact that there still exists an archaeology of gender can be seen as symptomatic of this issue. Certainly, there ought to be by now sufficient awareness of the complexity of the construction and maintenance of social identities to make gender an integrated aspect of how we study past societies, along with other variables, such as age, class and ethnicity. However, as we will see gender often remains unquestioned in ways which are detrimental to our overall understanding of past social construction: by assuming such a fluctuating and fluid variable can ever be assumed as a known entity without thorough investigation, we further the narrative of our own gender myths rather than attempting to comprehend past gender constellations.

The second is the ongoing challenge of presentism, which is especially pertinent when it comes to questions of gender and social norms: that the present influences the past is unquestionable, but equally the past influences the present, and can be used to naturalise and legitimise social inequalities. Thus, the idea that current Western gender roles of binary male vs. female values, behaviours and mind-sets are inescapable and natural is kept buoyant by the errant belief that this is a historically permanent feature of all societies, and this belief is kept alive to some degree by archaeological interpretations. This premise has seen extensive critique by archaeologists with an interest in gender over the past few years (Danielsson and Thedéen 2012; Fuglestvedt 2014; Ghisleni et al. 2016), as well as facing critique on a broader level through studies within the field of biology which question gender as binary (Fausto-Sterling 1993; Arboleda et al. 2014; Joel et al. 2015). Regardless of these challenges, it remains influential to the degree of being pervasive. We need to question why this narrative remains so strong by examining its role in the maintenance of our social cohesion. The belief in there being two ways of being, either male or female, is fundamental to the construction of modern western society (Connell 1987:64; Laqueur 1990; Butler 1993:3; Jordanova 1993:479). Challenging this means challenging the very fabric of our society’s collective understanding. However, this is nonetheless necessary if we wish to approach an archaeology where gender is not assumed a known entity.

The third point deals with the rise of intersectional perspectives, and is here presented as an avenue out of the current rigid system of gender which we tend to use when addressing the past, and into a more fluid and open way of interpreting social identities. This means understanding that social identities are constructed according to intersecting lines of discrimination and privilege (Crenshaw [1989] 2011), and that this will have a direct bearing on the interpretation of social groupings and their interactions.

Together, these points address what can be perceived as the current challenges in archaeological theory as related to gender issues today, whilst simultaneously proposing a different way of approaching a complex, and indeed potentially contentious issue.

Has Gender Gone Mainstream?

The first issue which must be addressed is whether or not gender archaeology has ‘worked’. It is not the aim of this article to provide a history of gender archaeology: for this, the reader is referred to any number of succinct publications detailing the origins and development of gender interest in archaeology (the recommended one being Gilchrist 1999). However, it makes a point in itself that the text recommended for getting to grips with gender archaeology is itself now two decades old, whilst the need to address our academic practices remains. This in itself tells us something about the success of gender archaeology so far.

The 1980s and the 1990s saw gender put on the agenda (Conkey and Spector 1984; Gero and Conkey 1991; Wylie 1991b; Nelson 1997; Gilchrist 1999) with a rise in theoretical developments and the realisation that gender was not a fixed or a given in any social setting. And yet it is fair to say that after the initial jolt which this caused, the general way in which archaeology portrays social order has not changed considerably. Thus, although there have been valuable contributions to the field in the last 20 years, mainstream archaeology cannot be said to have progressed past a token inclusion of gender in most cases. The current rise of intersectional perspectives has rejuvenated the debate on gender in archaeology somewhat, but this has not yet been sufficient to make gender a truly integrated aspect in studies of past social order. By this, whilst recognising the excellence of work done within this field, it is highlighted that there has not yet been a significant impact outside of quite a narrow academic sub-community. Nor has it, as of yet, succeeded in deepening the general understanding of gender past a perfunctory binary code, as will be further explored below.

Acceptance is Not Enough

A troubling trend in recent discussions of gender in archaeology is the belief that gender is now accepted in the mainstream. As was recently expressed in an article discussing gender configurations in the Neolithic: “Overall, gender archaeology in European prehistory continues to be widely accepted” (Robb and Harris 2018:130). And yet, acceptance is not sufficient, because the models within which we work remain rooted in ideals of gender as a binary either/or proposition and rarely question the ways in which we examine this in the past. Thus, even though more people arguably use terminology which appears to reference gendered aspects of social identities, this does not inevitably mean that our understanding of these issues has been advanced. Acceptance is indeed what we are currently facing: acceptance with the implicit understanding that it does not necessarily require agreement. What this means in effect is acknowledgement, without the necessary changes to interpretative frameworks. What is really needed is a widespread agreement and adoption of theoretical concepts and a deep and fundamental review of those methodological and theoretical frameworks which have facilitated the perpetuation of staid, binary gender constellations without further question.

Reviewing the Mainstream

It is also necessary to examine this idea of gender as mainstream and to ask how legitimate this claim really is.

To start with my current field of research, which is situated within Viking Age studies, a brief overview suffices to conclude that considerations of gender remain firmly on the margins in many cases. That more attention is paid now than what was common 40 years ago to the fact that the past not solely populated by wealthy men is certainly true, and yet it is clear that gender is still perceived as a known factor. It is, in some cases, easy to get the impression that gender is included more as an exercise in ticking the right boxes, rather than being integrated as a fundamental aspect of the construction of social and collective identities. Indeed, a brief survey of academic publications within my current field of Viking Age studies demonstrates two different ways of ticking the gender box:

  • The first is typically found in books dedicated to describing society and culture, where a common approach is to dedicate one chapter to women. Often this same chapter will also deal with family, as if women have a natural affiliation with family life that men lack (Roesdahl 1987; Sawyer 1992; Sawyer and Sawyer 1993; Sigurðsson 2010; Jesch 2015), a problematic assumption in itself.

  • The second approach is manifested in books which touch on concerns regarding women’s social status under any relevant headings and subheadings, and which then gather these references up in the index. These same indexes do not reference men and men’s activities, presumably as these are too numerous. The implication is that men are active as a norm, whereas women are active only in exceptional cases (Sawyer 1996; Christiansen 2002; Sigurðsson 2008; Williams et al. 2013), giving the impression of a society run in the main by men.

In Viking Age scholarship generally, there is a pervasive trend to cast men and women in fundamentally different roles, as we will return to below. The evidence for such a sharp division, however, is questionable. Indeed, a review of Viking Age social roles and pursuits shows that there are very few areas where both women and men could not participate (as discussed at length in my recent PhD thesis: Moen 2019), and yet the majority of publications do not take this into account when presenting a sharply gender-segregated society (for a few examples, see Christiansen 2002; Jochens 2002; Jesch 2015).

Moving away from Viking Age studies and over to other examples, a relatively recent article examined the impact of gender archaeology through the simple yet telling expedient of reviewing the proportion of journal articles dedicated to concerns of gender, with the resulting conclusion that gender archaeology has as yet to achieve a deeper impact beyond its ‘core constituencies’ (Tomášková 2011:118). A similar conclusion was reached independently by another scholar writing a year later with slightly different data (Danielsson 2012:229), thus showing that two separate review processes came to the same conclusion of gender remaining on the margins in archaeological research.

A recent article focused on demographic growth in the Upper Palaeolithic in south-western France has further summarised how gender is certainly not integrated into the mainstream of Palaeolithic studies (French 2018). Similarly, Bettina Arnold’s recent article regarding gendering mortuary contexts in the European Continental Iron Age has made it clear that this particular region and area do not yet deliver on the premise of gender as integrated into overall enquiries either (Arnold 2016).

It is important to make the point that this article is not meant to in any way detract from the many excellent works dedicated to questions of gender. It is instead intended to make it clear that it is still the case that we are looking at an academic landscape where things are either typically researched and written with a gender focus, or else they tend to bypass questions of gender or at best throw a cursory glance over what is a complex and fundamentally important part of social and collective identities. This article aims to highlight how far away from the mainstream gender archaeology still remains, and specialised enquiries into gender falls outside this. With a slight apprehension of sounding flippant, these studies preach to the choir in the sense that they appeal to fellow researchers with an express interest in gender, and are often overlooked by those without.

Though this is hardly a comprehensive overview, it is sufficient to show that the idea that gender is somehow ‘mainstream’ is an illusion.

The Illuminating Birka Case: A Quick Measure of Gender Awareness in Viking Age Archaeology

As a specific example of how gender can still prove controversial, it is worth outlining the reception, and resulting discussion, of a recent article detailing an aDNA analysis done on the skeleton of grave BJ581 from the Viking Age town of Birka in modern day Sweden, which revealed that the grave was that of a woman buried with a range of weapons and other artefacts traditionally associated with men in the Viking Age (Hedenstierna-Jonson et al. 2017; Price et al. 2019).

For anyone who has worked with women’s roles and gender ideology in the Viking Age, this ought not have come as a complete surprise, as women with weapons are known both from other burials and from written sources. Archaeological examples include numerous graves with axes at Kaupang (Moen 2011), the famous shield maiden at Nordre Kjølen in Norway (Mørck 1901), the discovery that two graves from Viking Age York contained women with weapons (McLeod 2011:345), a burial from Bogøvej, in Langeland in Denmark where a young woman was buried with a battle axe (Pedersen 2011:49). These are but a few examples out of many (see Gardela 2013 for further examples). Amongst written sources, Saxo mentions female warriors (Grammaticus [Approx. 1200] 1996), whilst several sagas refer to women fighters (Andrén 2008). The historian Johannes Skylitzes, whose works are partly preserved in Cedrenus’ history, details women warriors amongst the Rus forces from a battle in Bulgaria in AD 970 (Price 2002:332), and another example is the twelfth-century Irish chronicle Cordagh Ghaedhel re Gallaigh, which tells of a Viking attack on Dublin in the nine century, where one of the fleet commanders was known as “the Red Girl”, on account of her red hair (Clover 1993:366; Price 2002:74). There is also extensive pictorial evidence, in the form of small figurines depicting armed women, such the so-called Hårby Valkyrie, depicted below. Indeed, considering the amount of evidence, the suggestion that women may have at times been involved in warfare (as succinctly argued in Raffield 2016) seems like it ought to be fairly uncontroversial and not very surprising (Figure 1).

Figure 1
figure 1

Image of the Hårby Valkyrie. Photograph taken from Odense museum’s webpages. Attribution: Asger Kjærgaard, Odense Bys Museer

Full size image

However, the reception of this article demonstrated that it did come as a surprise for many. Media interest was high, and a number of prominent academics issued statements of support or disbelief.Footnote 1 Some went so far as to refuse the findings completely, others expressed doubt that military equipment necessarily makes a burial a warrior.

The first point can be countered on the grounds that the article in question presented a very thorough piece of research, which presents no cause to categorically refute or deny its results. A second article published in early 2019 has further elaborated the sound scientific basis of the analysis as well as providing further context regarding the find (Price et al. 2019) and also dealing with the reception to the initial article. As the authors’ state in the supplementary material, the accusation that they had somehow studied the wrong bones seemed to stem from a fundamental inability to accept the studies findings: “The underlying theme seemed to be that a female warrior was somehow a contradiction in terms, and therefore reasons must be found why our analysis was incorrect, even at the most basic level of having studied the wrong bones” (supplementary material for Price et al. 2019:6).

The second point, however, deserves consideration, as it is a valid concern that grave goods do not necessarily denote social status. However, if it is no longer to be the standard that weapons denote military rank or warrior status, we must go through the entire body of Viking Age burials and nuance the discussion for male and female burials alike. It must be admitted that the presence of weapons in male burials, though not necessarily equated with full warrior status, carries with it the academic consensus that the identities of the deceased were in some way affiliated with these items (as expressed in for example Jesch 2010; Schjødt 2011; Harrison 2015). The question here then becomes why perceived gender justifies a different interpretation. The answer can be suggested because it goes against what is recognised in academic circles as an inviolable line between male and female spheres, which may be more firmly cemented in what we can call accepted knowledge, than it is founded in the actual material.

Related to the above is the third point that needs to be examined in the way in which many expressed doubt of the article’s findings, because it did not fit with what they thought they knew. This point was made at times subtly and at other times more bluntly. The argument that this woman could not have been a warrior because women in the Viking Age were known not to be warriors is concerning to say the least, and it raises the question of what kind of science refuses evidence because it does not fit with preconceived ideas. As the authors of the original study stated in their most recent article: “it is not supportable to react only now, when the individual has been shown to be female, without explaining why neither the warrior interpretations nor any supposed source-critical factors were a problem when the person in Bj.581 was believed to be male” (Price et al. 2019:192), which sums up the issues latent in arguments which take issue with the potential implications of this burial. In short, we cannot now, once the body in question has turned out to be different from expectations, decide that the grave represents something completely different, at least not without compromising the integrity of archaeological interpretation of mortuary contexts from the Viking Age in general.

The ultimate contribution of the Birka burial BJ581 has been to restart an old debate by bringing new facts to the table. The findings of the article are irrefutable in the sense that there is physical evidence that a woman was buried with weapons and other artefacts indicating military rank. What is fluid and negotiable is not here the physical sex of the deceased; it is the interpretations which modern archaeologists impose upon it.

The Past in the Present: Gender Inequality as Legitimised Through Historicity

It may not be very in vogue to bring up the subject of presentism; indeed it is easy to get a sense that many researchers believe this has been discussed to exhaustion and beyond. However, I consider that when it comes to gender, this is an ingrained tendency with far-reaching consequences which has remained largely unexplored.

There remains a lingering assumption that gender is somehow female, or at least mainly the concern of women. This is interlinked with the above point of indexing women, where the ‘gender question’ is dealt with along with women and women’s roles. This betrays a conceptual misunderstanding of what constitutes a society, which is necessarily made up of a multitude of social categories, all creating a larger whole, where gender is but one factor.

The past is a product of the present, but it is also an active part of the present in which it is created. Thus, if we are told that gender inequality is historically inevitable, this lends support to the idea that men and women are fundamentally different and should be judged by different standards.

Gender as Inevitable and Natural: Naturalising Gender Inequality Through Prehistory

The weight of history is powerful, and by claiming that something has always been a certain way, gravitas and weight is added to a given argument. When it comes to social inequality, this can quickly become contentious and controversial. Archaeology as the study of the human past is incredibly powerful. Consider for a moment how Nazi Germany used archaeology to justify their ideologies of the master race (Arnold 1990). Consider also how our stories of ‘progress’ seem to prescribe a natural progression from primitive cultures to sophisticated societies (McGlade 1999; Lucas 2005; Dobres 2010; Maynes and Waltner 2012), and the subtle implications this bears for the ‘correct’ development of human societies.

The past matters, in other words. How we talk about the past, interpret it and present it, all these things matter, because ultimately these ideas will filter back into how we perceive our own society. Origin stories provide inevitability, absolving more recent history from blame when it comes to social injustice.

That archaeology still has a long road to travel before we can say that unconscious biases, shaped by our modern gender stereotypes are behind us, can be evidenced by a few examples:

In academia, the assumption of a universal, gendered division of labour still persists despite having been called out as an androcentric bias by pioneering gender archaeologists (Conkey and Gero 1991; Gero 1991; Gero and Conkey 1991; Wylie 1991b). To use Viking Age scholarship as an example again here, there is a widely assumed division of labour whereby women were responsible for work inside homes, whereas men got to take care of everything outside. This model is so common as to be ubiquitous, appearing everywhere from conference papers and presentations to academic publications (Roesdahl 1987; Solberg 2003; Callmer 2006:189; Sigurðsson 2008:44; Eriksen 2015:93 and 163). Closer examination, however, will reveal that it rests on assumptions more than on the material record and is traceable to nineteenth-century ideals (Arwill-Nordbladh 1998). This was highlighted over 20 years ago in Elisabeth Arwill-Nordbladh’s seminal investigation of the creation of gender stereotypes regarding the Viking Age (Arwill-Nordbladh 1998). This study is widely known and cited within Scandinavian academic circles, and yet the stereotype persists, presumably because it fits so well with our own ideas of the home as a female sphere, which are themselves a product of the nineteenth-century industrialised urbanisation process.

Turning to popular discourse, use of our ‘stone age ancestry’ is rife when it comes to explaining behavioural patterns, and comments on how men are better at certain tasks because of their evolution as hunters are all too common (see Fine 2016 for a dissection of this). This is problematic in a variety of ways, as it ultimately relies on the outmoded Man the Hunter paradigm (Lee and DeVore 1968). It assumes that only men hunted, which is by no means a certainty (Dewar et al. 2006). It also ignores the important socioeconomic role of gathering, traditionally assigned as the domain of women (Nelson 1997:85). But perhaps most importantly it relies on ethnographic analogy, a difficult and contested field of study at best (Gosselain 2016), and also one that is quite likely to carry forward the inherent androcentrism of early ethnography (as evidenced in some of the classical literature Lévi-Strauss 1969; Mauss [1925] 1990; Rubin [1975] 2006).

Finally on this subject, museums need to be mentioned: they are how many people first encounter the past, through school groups and family outings. Consequently, museums carry a responsibility to make sure that what they teach is correct as far as our knowledge goes, and does not carry implicit messages of exclusion. When it comes to gender, I argue that this responsibility is often not met. What many children will be told when they enter exhibitions is that men were active agents in the past, whilst women are few and far between, and when they occur are often associated with domestic activities (Gifford-Gonzales 1993; Bergsdóttir 2016). In other cases, women are simply removed from the story, as is the case in a new short film currently showing at the Viking Ship museum in Oslo. This film shows the travels and the eventual burial of a (male) Viking chieftain, using predominantly material from the (female) Oseberg burial, an appropriation of power symbols which is both frustrating and distorting.

We must acknowledge the impact this can have on those who visit these museums. If we teach the misguided assumption that since the earliest dawn of human history, women have tended babies and looked after the food whilst men have protected them, hunted and fought, we instil an expectation of this as natural and normal. From it being natural comes the logical extension that is also right, and from there it sets expectations of what girls and boys should grow up to be and do. This is not only limiting and potentially damaging, but considering the fact that we lack evidence to support these representations, it is indefensible.

Either/or: Gender as Binary or Gender as a Spectrum

Gender is a complex term. Arguably so complex that we ought perhaps not be talking about it at all without delimiting what we mean by it.

Early feminist scholarship tended to see sex and gender as clear and distinctive categories, where sex was a fixed biological category, whilst gender was seen as socially created and therefore culturally variable (Gilchrist 1999:9; Skogstrand 2006:110). In this view, gender is seen as culturally constructed, a physical manifestation of biological differences between men and women (Arwill-Nordbladh 2002:202), and a crucial question becomes to what degree gender is made or born (Gilchrist 1999:10).

Subsequent developments brought with it a move away from accepting sex and gender as neatly divided categories, or as inevitable expressions of physical differences. Rather, the realisation dawned that the two-sex model dominant in the modern West is by no means universal (Laqueur 1990; Nordbladh and Yates 1990:227; Arwill-Nordbladh 1998:64; Wiesner-Hanks and Meade 2004:2–3). This can be further put into context with the work of for example Anne Fausto-Sterling on the non-binary qualities of gender (Fausto-Sterling 1993), and with recent studies which have failed to identify universal sex-based differences in male and female brains (Joel et al. 2015). It can also be contextualised with re-examination of the history of biological arguments which show a profound bias in the way in which the results were interpreted, to the degree of invalidating many of the founding studies used to justify sex-differences as decreed by nature (see Fine 2016 for a comprehensive discussion of this). Together, these points provide a solid platform from which to argue that it is time to reconsider how we understand gender, and the tools with which we study it in the past.

This scepticism towards binary gender as the norm has also started to make its mark in archaeological theory, as manifested in a recent special issue of the Journal of Archaeological Method and Theory (Ghisleni et al. 2016) which dealt with the rejection of binary constellations, or as the authors’ termed it ‘binary binds’ (Ghisleni et al. 2016). Making the case that seeing gender as a simple either/or between masculine and feminine is limiting, and shapes our understanding and interpretations in far reaching ways, it is to be hoped that this study will have profound implications on the way in which gender is discussed in archaeology. It returns to the idea that gender is an expression of something natural, of biological sex, and that this is inviolable and a self-evident part of social construction. As long as this remains unchallenged, we will keep seeing interpretations of the past which reproduce our own gender constellations, limitations and roles. The questions we ask shape the answers we get. It is time to assess whether or not these questions are always the best ones.

Reappraising Gender and Limiting the Importance Accorded to It

Intersections, Overlaps and Different Levels of Discrimination

The last few years have seen a steep increase in the number of publications which fit under the, admittedly rather broad, umbrella of intersectional approaches (Villa 2011; Arwill-Nordbladh 2012; Danielsson and Thedéen 2012; Fahlander 2012; Thedéen 2012). And yet, intersectionality is hardly a new term, as it was first coined in 1989 by Kimberlé Crenshaw, specifically in relation to black feminism, and the intersecting lines of oppression which form a person’s experience (Crenshaw [1989] 2011).

Its introduction into the language of archaeological theory, however, has come about over the last decade or so, and must be judged a good thing in many regards. The term Old wine in new bottles (Vivar et al. 2011:2) has been, though in a tongue-in-cheek manner, applied to intersectionality in archaeology, as it is certainly not the case that earlier practitioners of gender archaeology were unaware of the fact that there are multiple levels on which a person’s identity is formed, and multiple ways in which a person can be discriminated against which shapes their overall experience. Indeed, this concretisation of how experience can be gendered, and how this can be relative to a host of other factors, such as age, religious persuasion, physical ability and skin tone, has been a cornerstone of gender archaeology more or less since its inception (as a cursory glance at some of the seminal works will demonstrate Conkey and Spector 1984; Conkey and Gero 1991; Gero and Conkey 1991; Wylie 1991b).

However, the new interest in intersectionality has brought with it an increased awareness of the multidimensionality of human experience, which fits in very well with the increasing stance against binary constructions of social identities. Identities are multifaceted, and experiences are shaped by inclusion and exclusion in part based on criteria such as gender, education, race and age.

Intersectionally informed approaches represent the only way we can truly see gender integrated into archaeological theory, by seeing beyond gender: not by disregarding it, and assuming it is no longer a necessary factor to consider when studying past societies, but rather, seeing that gender is but one part of a complex and composite image, and we ought not let gender stand in the way by ascribing it paramount importance over other integral parts of social identities.

Towards a Truly Gendered Archaeology

A number of things have to change in the way gender is ‘done’ in archaeological method and theory.

First, we need to stop treating gender archaeology as a sub-discipline. Gender needs to be a natural part of any study of past societies, and this means more than dedicating a single chapter or a few pages to ‘what women did’. We must instead work towards an understanding of a past populated by people, with the awareness that these people may have structured their world in quite a different way from what is the modern Western norm. Gender may, or may not have been, of paramount importance. The material in question should lay the foundations for such an interpretation, not our own experiences. When we gender burial assemblages for example, we unconsciously impose limitations on the material. We create expectations by assigning ‘male’ and ‘female’ status according to our own understandings of these concepts. I argue that a better starting point would be to try and avoid these restrictions and instead see what the assemblages can tell us across and within the groups which can be discerned on the basis of grave goods for example.

Second, awareness of presentism and the dangers therein needs to be fully recognised. Thus, if statements are to be made of what roles women and men carried out in the past, these need to rest on solid empirical bases, not on assumptions drawn from the researchers own experience or on past misrepresentations in earlier canonical texts.

Third, we need to address how we disseminate our information to the wider public. This is especially true as regards museums, but also through other forums such as increased open access publishing, widely disseminated blogs and public lectures.

If these three preliminary points are addressed, we may indeed see an approach to a truly gendered archaeology in the near future.

Gender archaeology will truly have worked when we get to a stage where gender is not seen as something that needs to be figured out so that we can fit prehistoric populations into their appropriate roles, but rather as something that needs to be understood as a component of social organisation, and not an answer in and of itself.