© 2018  All content copyright of Canine & Feline Behaviour Association. Click here for Company information Canine & Feline Behaviour Association Click to enlarge the CFBA logo Click to enlarge the CFBA logo
Home Members Join Us Legal Experts Vet Referral Information Media Contact Us
Click to go to the CFBA Facebook page Click to go to the CFBA Facebook page CFBA Magazine

We should not deny dominance in dogs

Dr David Sands FCFBA offers an insight into current canine behaviour science [Part 1] commencing with:

Dominance in domestic dogs revisited: Useful habit and useful construct?’

[Published online in April 2014

Matthijs B.H. Schilder, Claudia M. Vinke

(Department of Animals in Science & Society)


Joanne A.M. van der Borg

(Department of Animal Sciences, Behavioural Ecology Group, Wageningen University, Wageningen, The Netherlands)].

[My notes and comments are given in parenthesis]

[Science paper abstracts and quotes are given in italics]

Their Abstract is repeated here in full as follows]:

‘In the last decade, the validity and relevance of the dominance model was regularly put into question regarding relationships between canids like dogs and wolves, and consequently, human-dog relationships as well. The concept underlying this model, scientifically defined as an intervening variable reflecting status difference between individuals, is applicable when formal status signals symbolize the long-term relationship between individuals, resulting in a formalized dominance hierarchy.

This article reviews the basics underlying the concept of dominance and reflects on the value and importance of some new quantitative studies on the applicability of the concept of dominance in domestic dogs.

The conclusions are, first, that formal dominance is present in the domestic dog, expressed by context-independent unidirectional formal status signals. Consequently, formal dominance (e.g., submission) plays an important role in assessing status in dog-dog relationships.

Second, that nonverbal status-related communication in humans resembles that in dogs to a considerable degree, and hence dogs may be well able to interpret this human status-related nonverbal communication from their perspective. Dominance is therefore also likely to play a role in human-dog relationships. Hence, the dominance concept might be useful to explain the development of certain problems in dog-dog and dog-human relationships. However, enforcing a dominant status by a human may entail considerable risks and should therefore be avoided.’


© 2014 Elsevier Inc.

 [The line, ‘formal dominance is present in the domestic dog’ is significant because this stance reveals the Netherlands-based authors are in agreement with many responses - previously made available online by the CFBA and the PETbc* in relation to ‘Dominance in domestic dogs - useful construct or bad habit?’ Bradshaw, Blackwell, and Casey, (2009) and subsequently expanded in the publication ‘In Defence of Dogs’ by John Bradshaw (Allen Lane, 2011).]

[The researchers in the Netherlands state in their introduction that independent quantitative studies, published after Bradshaw et al, confirm the concept of dominance to be applicable in domestic dogs.]

[The authors cite three research papers, Cafazzo et al., 2010; Trisko, 2011; van der Borg et al., 2012 and indicate that these studies confirm the existence of the so-called formal dominance in dogs (my italics).

In addition they argue that formal dominance, ignored in previous discussions on dogs, is known in primates as a most important expression of submission or dominance. The authors then cite Bauer and Smuts (2007) as the exception because this research recognizes formal dominance in dogs, partly on the basis of a quantitative analysis of play-behaviour.]


[The paper by Bauer and Smuts is to be discussed in Part 2]


[Continuing from the introduction, the Dutch authors then lead into the main purpose of their research]:  

‘….the validity and power of the argumentations and data from authors opposing the idea of dominance playing a role in dogs and wolves and contrast these with the results from the recent studies using the Utrecht dominance model.’  

[The authors then indicate the subject of their final discussion]:

‘…the possible implications for the human-dog relationship’

[But not]


‘…the mechanisms of formation of dominance relationships and the connection between dominance and leadership.’

[The first section which follows the introduction is headed: ‘Some basics of dominance’. The second paragraph opens with - citing papers from 1974 and 1981]:

‘Dominance is regarded as an intervening variable that summarizes a set of behavioral differences between individuals’

[The various types of exchanges between dyads, animal A and animal B, -groups of behavioral elements that are performed with the same main directions within the dyads- are broken down into four for analysis quoted here]:

1) When main directions of submissive behaviors are reversed, these should correlate well with the behavioral measures reflecting dominance.

2) Main directions within dyads should be identical over different competitive contexts. That is to say that the main directions of behaviors within each pair of individuals in context A should be identical to those in contexts B and C.

[The example provided is: ‘the main directions of behavioral exchanges in mate-competition should be identical to those in food competition.’ An example of this type of consistency in dogs is given by the study of Cafazzo et al. (2010)’ -‘Effect of affiliative and agonistic relationships on leadership behaviour in free-ranging dogs’- to be discussed in Part 3]

3) Behavioral asymmetries should be stable over a considerable period to be measured over weeks, months, and even several years.

4) If a certain relevant behavior is found to be exchanged exclusively in one direction (animal A performs behavior X toward animal B, but [almost] never vice versa), it is said that such behavior functions as a formal dominance signal or even a meta-communicative signal (cf, van Hooff andWensing, 1987), conveying either dominance or submission.

[The four factors given are then followed by this sentence]:

‘There are formal signals that express dominance and formal signals that express subordination’

[This is cited in]:

‘Are primates behaviorists? Formal dominance, cognition, and free-floating rationales, Signe Preuschoft, 1999.’

[The Abstract]:

‘Contrary to a recent claim (D. Maestripieri, 1996), the concept of formal dominance (F. B. M. de Waal, 1986) is not dependent on higher order intentionality (D. C. Dennett, 1983). Instead, it is implied that primates have a concept of relative dominance and express this relational assessment by context-independent, unidirectional status indicators. Present evidence supports the view that primates are able to categorize social relationships. In a semiotic framework, a distinction is made between (a) ritualized displays that are symptoms of acute emotional states (e.g., fear) and (b) formal status indicators that are symbols for a long-term social relationship (e.g., subordination). Criteria to distinguish empirically between these functions relate to species membership and familiarity of interactants, consolidation of relationships, spontaneity of signaling, and specialization of signals.’

[Primate behaviour]

[The authors point out that ‘formal dominance signals’ have been described in several species including wolves and several primates citing van Hooff andWensing (1987) for wolves and de Waal (1977) for Java monkeys. ]

[The authors extrapolate on this point with the following]:


‘Formal dominance signals should be shown exclusively in the same direction, independent of the type of context. There are formal signals that express dominance and formal signals that express subordination (Preuschoft, 1999). In the latter case, these signals may also function to inhibit further aggressive or dominant behavior by dominants. ‘

[The primate example provided is given]:

‘…the ‘silent bared teeth display in Java and Barbary macaques is exclusively shown by submissive individuals toward dominants, independent of the context (de Waal and Luttrell, 1985; Preuschoft and van Schaik, 2000)’.

[The example in wolves is]:

‘…a high posture is shown by dominants toward subordinates, whereas a low posture is exclusively shown by subordinates toward higher ranking animals, again independent of the context (van Hooff and Wensing, 1987).’

[The authors then add]:

‘We will use the term status signals (cf, Flack and deWaal, 2004) as a shortcut, indicating signals that are involved in either signaling dominance or submission.’

[What follows is quoted directly from this section of the paper]:

‘In some articles (e.g., de Waal, 1989), an additional criterion to recognize formal status signals is added, namely that such a 1-sided signal should be shown not only during competitive conflicts but also outside conflicts, for example, during greeting. This has been shown to occur in free-ranging domestic dogs also (Bonanni et al.,2010)’.

[The Bonanni et al 2010 reference research to be discussed in Part 3]

‘The detection of formal status signals in a species is taken as proof that dominance relationships are important in structuring a given social organization. However, if formal status signals cannot be demonstrated, this does not exclude the possibility that dominance does play some role. Rather it suggests that its role might be far more limited (see Schilder, 1988).’

[Schilder 1988 cited is based on Captive adult Plains Zebra stallion research]

‘The degree of one sidedness by which a certain behavior is shown can be expressed in an index, called the direction (in) consistency index (van Hooff and Wensing, 1987).’

[The Hooff and Wensing is based on Java monkey primate research]

‘For example, a direction consistency index of 90% means that an average of 90% of all occurrences of that particular behavior in all dyads were shown by animal A toward animal B and only 10% by animal B toward animal A.’

‘For dogs, formal status signals were suggested by Bauer and Smuts (2007) in their article on play in dogs. They observed that “role reversals” occurred during play and involved chasing and tackles but never included mounts, muzzle bites, and muzzle licks, suggesting that these behaviors were invariant indicators of formal dominance during play in domestic dogs’.

[The authors placed ‘invariant indicators of formal dominance’ in italics].

[The research by Erika Bauer and Barbara Smuts, of the Department of Psychology, University of Michigan, Ann Arbor, published in 2007 is discussed in Part 2]

[The next significant series of statements in this section are]:

‘A last characteristic of dominance in a social group is that stable relationships often lead to a ranking of individuals, which is completely or nearly linear.’

‘However, triangles may exist, meaning that animal A dominates animal B, that animal B dominates C, whereas animal C dominates A. The degree of linearity is dependent on the number of blank relationships (no behavioral exchanges for that particular behavior in a dyad), the number of ties (both A and B perform a behavior to one another with the same frequency), and the number of triangles.’

‘Several procedures to construct rank order and test their linearity and steepness have been published (de Vries and Appleby, 2000; de Vries et al., 2000; Gammell et al., 2001; Bayly et al., 2006). Ranking individuals in a group may be useful to describe behavioral patterns in that group but may not (necessarily) reflect an important variable in social organization’

‘A quantitative analysis as described previously may lead to a variety of results: nonconcordant distributions of behaviors over dyads leading to nonidentical rank orders, or behaviors that are exchanged reciprocally, show an insufficient coverage, or a nonsignificant degree of linearity. An example of such an analysis leading to a limited applicability of the dominance concept as sketched can be found in the study by Schilder (1988) for plains zebra stallions. Therefore, the methodology is not bound to produce biased results but is precise and sensitive.’

[I read this sentence several times to understand the point…]

[The authors then cite three independent studies]

‘The fact that 3 recent independent studies on dominance in dogs using this methodology found formal dominance to be present in the dog cannot therefore, with any likelihood, be attributed to a bias inherent to the methodology.’

‘Finally, the validation of the concept of dominance must be found in its ultimate biological consequence, namely reproductive success. Indeed, the relationship between dominance status and reproductive success has been found in many species (see Ellis,1995 for a review; African wild dogs: Creel et al., 2007). Most, but not all these studies, confirm the biological importance of being high ranking, and therefore they point at the important role of dominance relationships.’

‘Especially in primatology, the study of dominance, its causes, and its biological implications has been developed to great depth. Many of these studies took many years, enabling assessment of (sometimes lifelong) reproductive output in relation to rank position. One such study has been performed on free-ranging dogs. The authors found that, although the pack comprised multiple breeding individuals, both male copulation success and female reproductive success were positively influenced by a linear combination of dominance rank, age, and leadership (Cafazzo, personal communication).’

‘Spotte (2012), on the other hand, summarizes that neither rank nor body size are important factors influencing reproductive success in the dog. Doubt is cast concerning this conclusion as, according to many articles, rank orders are not based on formal status signals.’

‘It has become clear that dominance relationships become established not because higher ranking individuals reinforce their status by being aggressive or showing formal dominance signals but because lower ranking individuals recognize supremacy by showing formal submissive signals (e.g., Rowell, 1974; Syme, 1974; van Hooff and Wensing, 1987).’

[The authors back up this ranking hypothesis or statement]:  

‘This statement also makes sense from a logical point of view: if acceptance is lacking, conflicts are bound to remain or individuals avoid each other and may leave their group (see Mech and Cluff, 2010 for wolves).’

[The authors then offer their key hypothesis outcomes in relation to domestic dogs]:

‘A state of repeated conflicts may occur when perceived asymmetries between 2 individuals are too small. In turn, this explains why conflicts between some dogs in the same household may be more frequent, persistent, and fierce than between other combinations.’

‘The probability of conflicts should be increased when dogs of the same sex, same age, and same breed are involved, leading to minimal asymmetry and hence, to more uncertainty with regard to their relationship.’

[Then arrive at the ‘assumed short-term function of dominance while pointing out their known studies, based on observations made in semi-captive wolves]:

‘With this point, we now arrive at the assumed short-term function of dominance, namely to restrain the number and severity of physical conflicts. In canids, we know of just 2 studies that show that a stable hierarchy decreases and an instable situation increases the occurrence of aggressive encounters in semi-captive wolves (Zimen, 1975; Moran 1982).’

[The previous citations are based on two captive wolf-group studies]

‘Nevertheless, it may occur that either the submissive animal does not recognize its status or the dominant animal will still reinforce its status, in spite of the submissive animal showing submissive behavior (Forkman and Haskell, 2004; Mech and Cluff, 2010).’

[The preceding statement is backed up in these publications]

[Forkman and Haskell - avian (hens) research]

[Mech and Cluff research is based on dominance behavior of Grey Wolves]

[However, the authors add a logical conclusion]

‘In principle, conflicts are costly because they consume energy and time, may lead to wounds and other risks, such as an increased vulnerability to predation as a consequence of a decreased attention toward predators.’

‘These are the reasons why individuals have to make an a priori assessment before entering into a conflict. This assessment would include processing information on whether the resource at stake is worth the risk, the chances of losing and winning, and the possible costs of entering into a conflict. The outcome of such a weighing should lead to a decision to fight or to refrain and retreat. Uncertainty concerning the resource holding potential (RHP) could lead to agonistic interactions. These provide individuals with opportunities to gain information on the strength of opponents, which consequently can result into avoiding fights with animals that could defeat them. In this way, physical harm or worse may be reduced (de Waal, 1989).’

[It is the RHP system that Bradshaw et al 2009 proposed as an alternative to using terms such as dominant and submissive in domestic dogs. Next, the authors dismiss the Casey summary (based on the Bradshaw et al study – of which she is a co-author) and cite: ‘Models of Dominance Hierarchy Formation: Effects of Prior Experience and Intrinsic Traits’ (J Beacham), their own, at that time, unpublished data and ‘The establishment of dominance relationships in a dog pack and its relevance for the man-dog relationship’ Netto, Van Der Borg and Slegers 1992 paper, based on a pack of 16 dogs in an ‘outdoor enclosure’, (Department of Comparative Physiology, University of Utrecht ).]

‘To form dominance relationships resulting in a rank order, animals do not need to have a concept of dominance, which is suggested by Casey (2009). It suffices that dogs know who is superior and who is subordinate, and a self-organized rank order emerges (Beacham, 2003). Moreover, unpublished data from Utrecht University showed that dogs also have insight into the relationships between third parties. If this were not true, interventions in ongoing interactions would be randomly directed. However, the data showed that they are clearly consistent with the rank order! Third-party interventions turned out to be more directed against lower than higher ranking individuals in an interacting pair and more against losers than winners (Netto et al., 1992).’

Also, trying to obtain a higher position does not need future planning, as Casey (2009) suggests. It just may happen provided there is some motivational factor that might be influenced by genetic (family) relationships, health, and environmental factors (see Chase and Seitz, 2011).’

[The cited reference ‘Chase and Seitz, 2011’ is given next]:

[‘Advances in Genetics, Volume 75, 2011, Pages 51–81, Chapter 4 – Self-Structuring Properties of Dominance Hierarchies: A New Perspective, Ivan D. Chase*, Kristine Seitz†

* Department of Sociology, Stony Brook University, Stony Brook, New York, USA † Department of Biology, Stony Brook University, Stony Brook, New York, USA]

 [The Abstract is here in full]:


Using aggressive behavior, animals of many species establish dominance hierarchies in both nature and the laboratory. Rank in these hierarchies influences many aspects of animals’ lives including their health, physiology, weight gain, genetic expression, and ability to reproduce and raise viable offspring. In this chapter, we define dominance relationships and dominance hierarchies, discuss several model species used in dominance studies, and consider factors that predict the outcomes of dominance encounters in dyads and small groups of animals. Researchers have shown that individual differences in attributes, as well as in states (recent behavioral experiences), influence the outcomes of dominance encounters in dyads. Attributes include physical, physiological, and genetic characteristics while states include recent experiences such as winning or losing earlier contests. However, surprisingly, we marshal experimental and theoretical evidence to demonstrate that these differences have significantly less or no ability to predict the outcomes of dominance encounters for animals in groups as small as three or four individuals. Given these results, we pose an alternative research question: How do animals of so many species form hierarchies with characteristic linear structures despite the relatively low predictability based upon individual differences? In answer to this question, we review the evidence for an alternative approach suggesting that dominance hierarchies are self-structuring. That is, we suggest that linear forms of organization in hierarchies emerge from several kinds of behavioral processes, or sequences of interaction, that are common across many different species of animals from ants to chickens and fish and even some primates. This new approach inspires a variety of further questions for research.’

[The Keywords listed]:

‘dominance hierarchies; social dominance; winner-loser models; prior attributes; jigsaw puzzle model; music notation; self-organization; transitive relationships; intransitive relationships; social cognition’

[The final statement in this section is]:

‘That many domestic dogs, if not most, do not show such an “ambition” may be also the result of domestication. In this respect, individual and breed differences are to be expected. Also, breeds may differ in their possibilities to show submissive and other behaviors (Goodwin et al., 1997), and breed-specific tail and ear postures may influence communicative (im)possibilities (Mertens, 2004; Leaver and Reimchen, 2008).’

[The first cited paper is ‘Paedomorphosis affects agonistic visual signals of domestic dogs’, Deborah Goodwin, John W.S. Bradshaw, Stephen M. Wickens, Anthrozoology Institute, University of Southampton]

[The Abstract is repeated here in full]:


Many of the structural modifications of modern breeds of domestic dog,Canis familiaris can be explained by changes in the rate of development, during domestication from the wolf,C. lupusThese changes have been dominated by paedomorphosis, or underdevelopment, so that the adult passes through fewer growth stages and resembles a juvenile stage of its ancestor. In this paper the effects of these processes on the signalling ability of 10 breeds selected for their degree of physical dissimilarity to the wolf are examined. The number of ancestral dominant and submissive behaviour patterns used during signalling within single-breed groups ranged from two (Cavalier King Charles spaniel) to 15 (Siberian husky), and this correlated positively with the degree to which the breed physically resembles the wolf, as assessed by a panel of 14 dog behaviour counsellors. When the signals displayed by each breed were grouped according to the stage of wolf development in which they first appear, those breeds with the smallest repertoires were found to draw most of their signals from those appearing before 20 days of age in the wolf, suggesting that physical paedomorphism has been accompanied by behavioural paedomorphism.’

[The second cited paper is ‘The Concept of Dominance and the Treatment of Aggression in Multidog Homes: A Comment on van Kerkhove's Commentary’, Petra A. Mertens (College of Veterinary Medicine, University of Minnesota , Journal of Applied Animal Welfare Science, Volume 7, Issue 4, 2004 – I have not been able to contact Dr Petra Mertens for a copy PDF as the email contacts given have expired]

[The final paper cited is: ‘Behavioural responses of Canis familiaris to different tail lengths of a remotely-controlled life-size dog replica’,

S.D.A. Leaver and T.E. Reimchen  (Department of Biology, University of Victoria, Victoria, British Columbia, Canada) Behaviour, Volume 145, Issue 3, pages 377 – 390, 2008]

[The Abstract is repeated here in full]:


‘The tail of dogs and allies (Canidae) is important for intraspecific communication. We used a life-sized dog model and varied the tail length and motion as an experimental method of examining effects of tail-docking on intraspecific signaling in domestic dogs, Canis familiaris. We videotaped interactions of 492 off-leash dogs and quantified size and behaviour of approaching dogs to the model's four tail conditions (short/still, short/wagging, long/still, long/wagging). Larger dogs were less cautious and more likely to approach a long/wagging tail rather than a long/still tail, but did not differ in their approach to a short/still and a short/wagging tail. Using discriminant analyses of behavioural variables, dogs responded with an elevated head and tail to a long/wagging tail model relative to the long/still tail model, but did not show any differences in response to tail motion when the model's tail was short. Our study provides evidence that a longer tail is more effective at conveying different intraspecific cues, such as those provided by tail motion, than a shorter tail and demonstrates the usefulness of robotic models when investigating complex behavioural interactions’.

[The next section from the Dutch authors is titled ‘Dominance in dogs: considering the pro and contra arguments in more detail’]

‘In this part, we follow the main arguments that would invalidate the concept of dominance for domestic dogs and comment on findings and conclusions.’

[The opening begins with a question: Personality trait or dimension of relationship?]

[The authors then draw attention to more research]:

‘Some authors claim that dominance is not a trait, but a characteristic of a relationship (Bernstein, 1981; Langbein and Puppe, 2004; Bradshaw et al., 2009). Consequently, ‘use of the expression “dominant dog” is meaningless’ (Bradshaw et al., 2009, p. 138). Comparable statements can be found in psychology. For example, Burgoon and Hale (1984) define dominance and submission as fundamental dimensions of personal relationships, whereas others recognize dominance as a personality trait (Kalma, 1991; Gangestad et al., 1992; Zebrowitz and Collins,1997). There has thus been considerable debate in human and dog personality research on how to score and interpret certain traits and over the classification of certain behaviors within trait spectrums (e.g., Jones and Gosling, 2005; DeYoung et al., 2013). For instance, submissiveness has been classified as a trait in its own right (Jones and Gosling, 2005) and also as a subtrait of neuroticism (Ley et al., 2008).

[I have highlighted the following sentences from the continuing discussion]:

‘There is no logical reason why a trait like dominance/submission would not be present in canids and other social species.’

[The linking sentence states]:

‘Data from studies on dog personality up to now were not linked to data on dominance in social groups.’

‘An attempt was made by Akos et al. (2014), who investigated leadership, dominance, and personality in a small group of dogs. These authors state on page 2 that “leader/dominant dogs have a unique personality: they are more trainable, controllable and aggressive, additionally they are older than follower/subordinate dogs.” ’

[The cited research is]:

‘Leadership and Path Characteristics during Walks Are Linked to Dominance Order and Individual Traits in Dogs’

by Zsuzsa Akos 1.*, Robert Beck 1., Mate´ Nagy 1,2,3.,

Tama´s Vicsek 1,2, EnikoKubinyi 2,4*

1 Department of Biological Physics, Eo¨tvo¨s University, Budapest, Hungary,

2 Statistical and Biological Physics Research Group of the Hungarian Academy of Sciences, Budapest, Hungary,

3 Department of Zoology, University of Oxford, Oxford, United Kingdom, 4 MTA-ELTE Comparative Ethological Research Group, Budapest, Hungary

- research to be discussed in Part 4]

[The Abstract is repeated here in full]:


‘Movement interactions and the underlying social structure in groups have relevance across many social-living species. Collective motion of groups could be based on an ‘‘egalitarian’’ decision system, but in practice it is often influenced by underlying social network structures and by individual characteristics. We investigated whether dominance rank and personality traits are linked to leader and follower roles during joint motion of family dogs. We obtained high-resolution spatio-temporal GPS trajectory data (823,148 data points) from six dogs belonging to the same household and their owner during 14 30–40 min unleashed walks. We identified several features of the dogs’ paths (e.g., running speed or distance from the owner) which are characteristic of a given dog. A directional correlation analysis quantifies interactions between pairs of dogs that run loops jointly. We found that dogs play the role of the leader about 50–85% of the time, i.e. the leader and follower roles in a given pair are dynamically interchangable. However, on a longer timescale tendencies to lead differ consistently. The network constructed from these loose leader–follower relations is hierarchical, and the dogs’ positions in the network correlates with the age, dominance rank, trainability, controllability, and aggression measures derived from personality questionnaires. We demonstrated the possibility of determining dominance rank and personality traits of an individual based only on its logged movement data. The collective motion of dogs is influenced by underlying social network structures and by characteristics such as personality differences. Our findings could pave the way for automated animal personality and human social interaction measurements.’

[I would welcome reader comments]

[A quote from the Dutch authors]:

‘These authors state on page 2 that “leader/dominant dogs have a unique personality: they are more trainable, controllable and aggressive, additionally they are older than follower/subordinate dogs.” ’

[The Dutch authors then make the following suggestion which is supported by cited research]:

‘Most likely dominant dogs will be of the “bold” type as defined by Svartberg (2005). This behavioral tendency often cannot be demonstrated at an early age in dogs or wolves (dogs: Beaudet et al., 1994; Diederich and Giffroy, 2006; wolves: Packard, 2003), although in a recent meta-analysis Fratkin et al. (2013) found substantial consistency for the dimensions aggression and submission in puppies, just as Svartberg et al. (2005) found consistency in personality in adult dogs.’

[The Svartberg et al paper was made available online in November, 2004 and was published in 2005 in Animal Behaviour, Volume 69, Issue 2, February 2005, Pages 283–291,

‘Consistency of personality traits in dogs’, Kenth Svartberg*, Ingrid Tapper†, 1, 2, Hans Temrin*, 1, Tommy Radesäter*, 1, Staffan Thorman‡, 3

(* Department of Zoology, Stockholm University, Sweden, † Swedish Working Dog Association, Sweden, ‡ Swedish Museum of Natural History, Sweden)]

[The Abstract is repeated here in full]:


‘We investigated the consistency of behaviour over repeated tests in dogs,  Canis familiaris. Dogs were tested three times, with an average of 30 and 35 days between tests. The behavioural test used in the study included 10 subtests that exposed dogs to various situations, such as the appearance of an unfamiliar person, play, prey-like objects, metallic noise and a suddenly appearing dummy. Studies using the same test with many dogs have revealed five specific personality traits, labelled Playfulness, Chase-proneness, Curiosity/Fearlessness, Sociability and Aggressiveness, and one higher-order, broader dimension, interpreted as a shyness–boldness continuum. We used these traits in the present study. We found significant correlations over the test series in all the specific traits as well as in the Boldness dimension. The magnitude of trait scores for Playfulness, Chase-proneness and Sociability, as well as for the Boldness dimension, was stable between tests. The scores for Aggressiveness and Curiosity/Fearlessness, however, differed between the first two tests: the intensity of behaviour related to fear and aggression decreased from test 1 to test 2, but the intensity of exploratory behaviour increased. This result indicates that these two traits in dogs are sensitive to novelty, although individual differences are also maintained in non-novel situations. The results suggest that playful, social, exploratory, avoidant and aggressive behaviour in dogs is influenced by stable dispositions; i.e. personality traits, that seem to have been important during the evolution of the domestic dog’

[The Dutch authors add]:

‘There are indeed dogs that show high postures toward many other dogs, just as there are dogs that show submissive behaviors toward many other dogs. This may reflect respectively dominant and submissive personalities, which may be recognized by many other individuals, although bold (certain) and shy (uncertain) would also be appropriate labels.’

[What follows are discussions and questions surrounding the validity-suitability of wolf behavior comparisons which ends in the following sentence]:

‘Most data on dominance in wolves stem from groups in captivity, where group composition may be abnormal and where levels of competition and aggression may be higher because of the lack of opportunities to disperse as wild wolves have.’

[The final (and significant) paragraph offers cited research for the debate and this is preceding Table 1 in which a summary is provided with comparisons detailed: ‘Signals related to (formal) dominance in wolves, dogs, and humans’]

‘Further arguments that the wolf model cannot be applied to dogs are first, that dogs do not live in wolf-like cooperative packs and second, that dog behavior differs radically from wolf behavior because of domestication (Bradshaw et al., 2009). The studies by Pal et al. (1998) and Pal et al. (1999) are cited by Bradshaw et al. to illustrate these differences (see later). More recent, however, a clear pack structure in free-ranging dogs has been revealed (Bonanni et al., 2010; Cafazzo et al., 2010, 2012). Although not all feral dogs may adopt a wolf-like pack structure, the data in the newer articles bear out that communication regarding status does not differ so much between wolves and dogs at all.’

[This significant information comes with the following caveat]:

‘Although the behavior of captive wolves may differ quantitatively from that of free-ranging wolves related to the level and intensity of competition, so far, the available data show that the principles of communication are identical. Because formal status signals in wild wolves do occur, dominance must also play a role there. That feral dogs do not always live in clear pack structures might complicate discovery of dominance relationships but does not inherently makes them nonexistent.’

[I attended an ‘inter-bitch aggression discussion’ at a conference where the significance of aggression between female domestic dogs was that feral adult female dogs hold a ‘similar rank’ and defer to the breeding female which has alpha status]:

‘In feral dogs, reproduction is not linked with dominance. Pal et al. (1999, 2003, 2005) could neither find clear displays of dominance or submission in their group of feral dogs nor frequent aggressive encounters between females.’

[The Dutch authors discuss the Pal et al research on feral dogs]:

‘They found little reproductive suppression in females, and they found that infanticide only occurred rarely. However, Pal et al. (1998) established dominance hierarchies based on aggressive encounters, which is supposedly not an adequate measure, as discussed before. Moreover, Pal et al. (1999) could not detect a relationship between dominance and the number of matings. From this, Bradshaw et al. (2009) concluded that “reproduction in feral dog groups appears not to be controlled by a wolf pack type of dominance hierarchy.” It can be questioned if such a conclusion can be made based on data on this small group of animals; and it certainly should not be generalized. It remains also unclear from their article if Pal et al. (1998) have investigated whether asymmetries in behavioral exchanges are concordant before lumping behaviors into groupings for investigating dominance and its consequences. In a later study, Pal (2010) refers to Cafazzo et al. (2010) regarding their defining criteria for dominance in litters of feral dogs: he now used “upright and stiff body postures with head held high and ears pricked, and growling and baring teeth.” In this study, he states that dominance hierarchy develops at an early stage of life and that “female and male domestic dogs form intra-sexual dominance relationships”

‘Pal’s findings might be simply a matter of a small data set, data arrangement, and analysis. As mentioned before, a recent study by Cafazzo et al. (personal communication) has shown a relationship between dominance and reproductive success in free-ranging dogs.’

[The Dutch authors then turn to the Bradshaw research]:

‘Dogs are not likely to strive for dominance. Bradshaw et al. (2009) and Casey (2009) presented some data of their own research on a group of neutered male dogs. In this study, no clear rank order could be found, but they did find some “consistent dominant or subordinate relationships” within that group. Based on these findings, they concluded that “it now seems unlikely that interactions between dogs are always, or indeed ever, driven by the aim to achieve status within a social group.” ’

[The significant sentence follows and confirms the thoughts of many of us that have issues with Bradshaw’s et al findings]:

‘This conclusion is based on the investigations of a group of castrated individuals of the same sex, seemingly lacking any competitive incentives (e.g., females, food or other highly valued objects). In such a situation, there is not much use to achieve a higher rank.’

[The final section by the Dutch authors is titled ‘Dominance and the human-dog relationship’ and the significant sentences are as follows]:

‘The 3 recent quantitative studies demonstrate that a limited number of formal status signals shown by dogs indicate dominance relationships (Table 1). They confirm the idea of Bauer and Smuts (2007) that formal dominance in dogs exists.’

‘McGreevy et al. (2012) give an interesting overview of (dis) similarities between dogs and people regarding communicative signals. They mention a number of behaviors (Tables 4, 5, 6: p. 110-112) like stand over, stare, place paw on to forequarter, averting eye contact, submissive grin, and submissive posture as signals that occur in humans and dogs.’

[McGreevy et al. (2012) research is by an Australian group from various institutions or workplaces including the  Faculty of Veterinary Science, University of Sydney, Sydney, NSW, Australia;Research Services Division, Deakin University, VIC, Australia; Guide Dogs Victoria, Kew, VIC, Australia; Anthrozoology Research Group, Monash University, VIC, Australia; and Behaviour Counselling Service, Oakleigh South, VIC, Australia. The authors are: Paul D. McGreevy, Melissa Starling, N. J. Branson, Mia L. Cobb, Debbie Calnon, ‘An overview of the dog–human dyad and ethograms within it’.]

[This research remains pending for my review but the Abstract offers an insight]:


‘This article reviews the literature on the complex and variable nature of the dog–human dyad and describes the influence of terms such as ‘‘dominance’’ on attitudes that humans have toward dogs. It highlights a legacy of tension between ethology and psychology and notes that some practitioners have skills with dogs that elude the best learning theorists. Despite the widespread appeal of being able to communicate with dogs as dogs do with one another, attempting to apply the intraspecific dog ethogram to human–dog and dog–human interactions may have limited scope. The balance of learning theory and ethology on our interactions with dogs is sometimes elusive but should spur the scientific community to examine skills deployed by the most effective humane practitioners. This process will demystify the so-called whispering techniques and permit discourse on the reasons some training and handling techniques are more effective, relevant, and humane than others. This article explores the mismatch between the use of nonverbal communication of 2 species and offers a framework for future studies in this domain. Technologies emerging from equitation science may help to disclose confusing interventions through the collar and lead and thus define effective and humane use of negative reinforcement. The case for a validated intraspecific and interspecific canid ethogram is also made.

 2012 Elsevier Inc. All rights reserved.’

[The revealing sentence (after discussions focusing on comparisons - communication signals – human-dog dog-dog and whether they are related to dominance and submission) is]:

‘Whether humans are able to interpret dog behavior correctly remains to be seen.’

[Another area of research, available online, is cited in the final section in discussions]

Kujala MV, Kujala J, Carlson S, Hari R (2012)

Dog Experts’ Brains Distinguish Socially Relevant Body Postures Similarly in Dogs and Humans.’  

[This will be discussed in Part 5]

[The Abstract in full is]:

‘We read conspecifics’ social cues effortlessly, but little is known about our abilities to understand social gestures of other species. To investigate the neural underpinnings of such skills, we used functional magnetic resonance imaging to study the brain activity of experts and non-experts of dog behavior while they observed humans or dogs either interacting with, or facing away from a conspecific. The posterior superior temporal sulcus (pSTS) of both subject groups dissociated humans facing toward each other from humans facing away, and in dog experts, a distinction also occurred for dogs facing toward vs. away in a bilateral area extending from the pSTS to the inferior temporo-occipital cortex: the dissociation of dog behavior was significantly stronger in expert than control group. Furthermore, the control group had stronger pSTS responses to humans than dogs facing toward a conspecific, whereas in dog experts, the responses were of similar magnitude. These findings suggest that dog experts’ brains distinguish socially relevant body postures similarly in dogs and humans.’

[The summary given by the Dutch authors is]:

‘In conclusion, the notion of dominance in dogs in our opinion is a useful concept. It explains an important aspect of the relationship between dogs and also between dogs and humans. We want to emphasize that stability in a relationship does not arise from exerting dominance (over dogs by humans by using forceful acts) but by showing formal submissive signals (by the dog toward humans). Teaching a dog to accept humans as dominants and to accept related behavioral limits should be different from teaching appropriate behavior like sitting or lying down on command. The former depends heavily on learning to accept a submissive status, which chiefly necessitates an adequate socialization of the dog and a clear and consistent behavior by the owner. The latter involves reward-based training, which is being facilitated by a clear dominance relationship but not dependent on it.’


[My final thoughts regarding this extensively researched review surrounding the suitability of the use of the terms Dominance and subdominant in dogs are still to be established.

Extrapolating most of the supporting research in this paper, while interesting, is time-consuming. Perhaps it is less so when a reader is already working in a science department within an institution. With this in mind I thank those researchers who made, and continue to make, their work available free online.

This paper is important to those of us working with companion dogs because Matthijs Schilder, Claudia Vinke and Joanne van der Borg support our counter argument to the ‘Bradshaw standpoint’.

Post-reactions to Bradshaw have filtered into our work, as professional dog trainers and behaviourists, suggesting the current science argues we dismiss the terms of ‘dominant’ and ‘sub-dominant’ together with the concept of ranking in domestic dogs. Those among us who have developed a rational thought-process that, while it may not be useful to directly compare captive-wolf research -or even free-ranging wolf research- with domestic dog behaviour, it is not realistic either to dismiss the comparison altogether.

The domestic dog is a sub species to the wolf.  The time-scale effect of thousands of years of continuing domestication of dog breeds we keep today does make direct comparison between dogs and wolves a challenge. It is not unreasonable to examine parallel behaviours observed in wolves and domestic dogs We accept that there exists a practical difficulty for researchers to observe and map non-captive, free-ranging wolves and dogs over each season and many years. It is also impossible to directly compare them because the evolution of behaviour is not ‘fossilised’.

In terms of domestic dogs, no amount of research on ‘artificial dogs groups’ without controls for owner-influence, age, gender, diet, social and environmental experiences including ‘entire’ over ‘neutered’ dogs, can be deemed complete.  Researchers will continue to debate for or against the use of specific terms because they are left only with the compromised option of turning terms into subgroups of words to include the ‘scientific language and meaning’, in this case, the use of dominance and its opposite submission. We are left only with the ability to use dilutions of the use of ‘dominant and sub dominant’ when describing a dyad between two domesticated dogs. This leads us down the road of agreeing which ‘meaning of the words’ can be used in describing behaviour and questioning whether they should be applied only to people or also to other animals.

An example of debatable word-usage in dogs is the term anxiety in people and companion animals. In its use related to people, the feeling of anxiety or an ‘anxious state’ can be freely discussed with the amount of detail depending on the extent of vocabulary. The only words that are understood by dogs are those learned through phonetic-association which helps to form a communication-bridge between our two species that is reinforced by repeated actions.

An example of terminology confusion is the term anxiety a word used from the 19th century – derived from Latin – employed by psychologists and psychiatrists to describe our feelings of unease, worry and fear and the emotions and the physical sensations. These we might experience when we are worried or nervous about something. Anxiety is commonly used by veterinary professionals to describe a dog that presents displaced-behaviour and stress when separated from an owner. The semantic viewpoint suggests its use related to dogs is incorrect. We know the dog cannot describe or explain that it is ‘destructive’ or repeat-barks and howls or the eliminated faeces and urine puddles occur because it feels solitary confinement in its house-territory worrying and stressful. Most of us know what these behaviours indicate if anxiety is used to describe the home-alone ‘state’ and why a dog finds separation from its owner difficult to accept.

The Netherlands based authors’ research summary is sufficiently cited to confirm what professional dog trainers and behaviourists already have accepted, which is that some companion dogs display behaviour of a dominant-like demeanor and other dogs may be submissive or defer to them. The latter dogs are challenged by the dog perhaps instinctively using what may be the Resource-holding potential to avoid potential injury by either by growling or standing upright (or over them) with eye-contact and tail-fixed straight. The dominant dog may or may not then choose to attack and bite and behave aggressively. Those dogs that defer are more likely to lower their body, tail-wave side to side sometimes quite vigorously (wagging) or whine. They may have learned it better to defer than risk injury.  The dogs that do not defer may challenge so offer the same dominant-like behaviour in return.      

The scientists-academics have a professional agenda to develop theories, to create and stimulate research; to go macro, collect data and then use statistical analysis as proof for the variables, the lack of controls, the modeling and the minutiae.

Scientists-academics are no different than us; they are selective with information they include or exclude and in which publication they publish and eventually, which research they challenge or support and expand. Scientists choose which research they cite in agreement or to counter and any research they ignore and choose not cite. They have a life to live and work like most of us. Doing it well in the scientific world would be, in our opinion, to make science accessible and, through the use of clear language, understandable and therefore useful as information. This is way that we, the majority, can learn and gain knowledge.

My slight concern is that we are sending exploratory science instruments to the planet Mars and yet we still know little about human condition and the natural and unnatural world around us.]

[© Dr David Sands and the CFBA, February, 2016]

Magazine DVD Films Press Dog Behaviour Videos

Denying Dominance in Dogs is Not Helpful - Dr. David Sands