Playful teasing and the emergence of pretence

The study of the emergence of pretend play in developmental psychology has generally been restricted to analyses of children’s play with toys and everyday objects. The widely accepted criteria for establishing pretence are the child’s manipulation of object identities, attributes or existence. In this paper we argue that there is another arena for pretending—playful pretend teasing—which arises earlier than pretend play with objects and is therefore potentially relevant for understanding the more general emergence of pretence. We present examples of playful pretend teasing in infancy before and around the end of the first year, involving pretend communicative gestures, mis-labelling and almost non-compliance with prohibitions. We argue that the roots of pretence not only lie earlier in human infancy than generally acknowledged, but also are rooted in playful emotional exchanges in which people recognise and respond to violations of communicative gestures and agreements.

1 3 of less interest in terms of the pretending involved. Construing playful teasing as suppressed hostility, as earlier anthropological and psychological literature tended to do (Radcliffe-Brown, 1940;Burk, 1897), casts the process as an internal psychological construct rather than one which is co-constructed by the participants (Mills & Carwile, 2009), differs in prevalence across communities (Göncü, Mistry & Mosier, 2000) and means different things in different socio-cultural contexts (Göncü & Vadeboncoeur, 2016). Playful teasing can not only be prosocial, leading to intimacy and solidarity (Dynel, 2008;Loudon, 1970;Pawluk, 1989;Haugh, 2016;Haugh & Pillet-Shore, 2018), but can be motivated by a need for play and challenge-an 'incident affinity' as Nakano (1996) calls it-or the desire for creating events which up the ante in terms of relationships. Playful teasing is an altogether richer phenomenon for understanding pretending.
The link between playful teasing and pretending appears to be only partially-i.e., one-sidedly-acknowledged in psychology: conceptual analyses of playful teasing acknowledge the role of pretence, but conceptual analyses of pretence do not generally discuss playful teasing. Analyses of playful teasing often invoke non-literality which is an explicit identifier of pretence. The combination of an act in its literal or serious form and 'off-record' markers of non-literality or playfulness, is seen as characterising playful teasing (Keltner et al., 2001). Others have argued that playful teasing can be seen as 'joint pretense' (Clark, 1996, cited in Keltner et al 2001 or 'social pretense' requiring the ability to assume pretend roles and stances and interpret acts of pretence (Heerey et al., 2005). Empirical studies of playful teasing often use explicit categories of pretending in their operationalisation of teasing: 'mock challenges', 'mock commands', 'mock threats' (Keltner et al., 2001), 'pretending to fight' (Labrell, 1994), or 'jocular pretence' and 'jocular mockery' (Haugh, 2016).
However, studies of pretend play have generally paid little attention to playful teasing and have tended to ignore its implications for the emergence of pretend play in infancy 2 . Operational definitions of pretend play focus almost exclusively on pretending about the identities, attributes and existence of objects. To consider the significance of playful teasing for the emergence of pretend play it is not sufficient to assume that pretending with object meanings simply transfers to, or is enacted in engagements with, people. The two are importantly different. First, in order to tease you need another consciousness-the pretending in it needs to affect another person, fooling them or drawing them into the play. Second, playful teasing unfolds in meaning-it needs participation for its meaning to be constructed and evolve. The neglect of playful teasing in attempts to understand the emergence of pretend play in development is therefore seriously problematic.

3 2 Developmental research on pretend play
Almost by default, the study of the initial emergence of pretend play appears to have slipped into a definition of pretending that is limited to what children can do with toys and everyday objects. Most investigations of early pretending tend to use three operational criteria-substituting one object for another, attributing features or properties to things that are not really there and pretending the existence of non-present objects (Leslie, 1987;Lillard, 2017;Nielsen & Dissanayake, 2004;Thompson & Goldstein, 2019;Weisberg, 2015). These manipulations of object functions and meanings succeed the ability to use toys or objects according to their intended function (functional play) and develop progressively. Using an object as if it was something else appears at around 12-18 months (Li, Hestenes, & Wang, 2016;McCune, 1995), followed by first the ascription of pretend properties or animation between them and then pretending an object is present when it isn't 2017;Weisberg, 2015). Leslie (1987) contends that whilst functional play with real objects or miniature versions of everyday objects, such as pushing a toy car along the ground, or putting a spoon to the mouth of a doll (relating two associated objects together), simply demonstrates a conventional understanding of objects involving only a first-order representation, pretend play requires the "decoupling" of mental representations from "reality". In other words, to engage in pretend play he suggests that a child needs to be able to treat an object as if it were something else (e.g., a shoe as though it were a boat) whilst at the same time appreciating that it is not really the same as the thing they are pretending it is. This equating of early pretence with the ability to manipulate object meanings and functions is therefore not seen as merely incidental but as chronologically and developmentally logical and central: "Evidence is consistent from all of the studies reviewed that initial pretending depends on knowledge of the functions and structures of real objects" (McCune-Nicolich, 1981, p. 76).
This focus on objects excludes a number of things from the developmental 'frame': the role of other people-particularly caregivers-in pretend play, playfulness within pretending with objects, as well as pretend actions with people. Also neglected, of course, has been early deception such as fake crying (Reddy, 1991;Nakayama, 2010Nakayama, , 2013; although it is not play, the pretence in fake crying is very relevant to the emergence of pretence. Why has this exclusive focus on object meanings developed in developmental psychology, with the almost complete neglect of the pretence contained in playful teasing using gestures, expressions or understandings? Part of the reason might lie in a theoretical shift in the last half century or more, from an earlier phenomenon-based and gradualist take on the emergence of pretending towards a more categorical and either-or stance; a shift also paralleled by one from naturalistic observations to experimental studies. Piaget's approach to pretending, for instance, was focused on observations of behaviour and included actions such as sleeping, drinking and even offering. Despite his commitment to a stage theory, he saw pretending as a gradual emergence in which the most basic form of the symbolic schema was the reproduction of an action pattern "outside its context and in the absence of its usual objective", a de-contextualisation later in development "applied to new and inadequate objects" (Piaget, 1962). In contrast, more recent theorists have sought to draw sharper and more internalist lines (Leslie, 1987;cf. Williams et al., 1 3 2001) and have distanced pretence as a representational ability from its contexts of occurrence. Distinguishing between functional play and pretend play, Leslie, Baron-Cohen and others portray pretend play not as elaborations emerging from simpler forms of play (Robinson, 2019) but as being the result of a completely different kind of ability-decoupling-which "once having emerged, does not develop any further" (Leslie, 1987, p. 420). Somewhere in developmental psychology's theorising about pretence, pretending has become a categorically distinct form of action, separate from its developmental history as well as its present contexts, originating in thought and limited to playing with sets of toys.
Another part of the reason for this focus on objects, may have emerged, embarrassingly, as a convenience-option: "…(pretend play) is often easiest to see with object-substitution pretence, in which one object is used as a stand-in for another: the classic banana-as-telephone scenario, or cases in which a child interacts with a stuffed toy or doll as if it were a baby" (Weisberg, 2015, p. 250). In contrast, playful teasing is a much more context-sensitive activity, not easy to elicit on demand with the result that empirical data are hard to obtain (Reddy, 1991(Reddy, , 2008. As Lillard (2017) notes in relation to empirical research on play fighting, there is a daunting difficulty with creating the right conditions and the right adversary, in order to elicit this (interpersonal) phenomenon. But the convenience-option has insidiously become a criterion and an empirical filter.
There may be a third and deeper reason for the neglect of (and lack of data about) interpersonal pretending in the developmental literature-a persisting belief in the relative irrelevance of the social. Piaget's neglect of the social contexts of pretending allowed the assumption that pretend play would emerge regardless of the presence of anyone to play with and uninfluenced by parental preferences. This led to the prediction that it was "unlikely that parents play …. (or) model such games" (Fein, 1981), an assumption strongly challenged by studies documenting the huge amount of parental involvement in children's pretend play (Haight & Miller, 1992;Garvey, 1982) and by variations across cultures in what play opportunities may be afforded to children by parents (Göncü, Mistry & Mosier, 2000). Imaginative play is fundamentally a socio-cultural activity and cannot be abstracted from the practices of communities (El'Konin, 1966;Göncü & Vadeboncoeur, 2016; see also Rossmanith, Costall, Reichelt, Lopez & Reddy, 2014). However, where studies do explicitly focus on social interaction in pretend play it is usually with reference to social interaction surrounding the pretend play, such as engaging with peers while also engaged in some form of pretend play or involving others in object substitutions. Teasing and joking don't really get a look in.
The neglect of the interpersonal in children's pretending seems also, perhaps not surprisingly, to be accompanied by the neglect of pretend play in adults. The operationalisation of pretend play in terms of activities using signifiers (e.g., an alternative object such as a piece of cloth) to represent the meaning of a signified (the actual object, such as a wig) may have led to a complete neglect of pretend play in adults (such as teasing or improvisation theatre) since there is no doubt about adults' ability to represent symbolically (Göncü & Perone, 2005).

3 3 Playful teasing in early development
References to playful teasing in childhood are evident in early writing (Groos, 1896(Groos, /1976Valentine, 1942) as well as more recent studies Miller, 1986;Eisenberg, 1986;Warm, 1997;Keltner et al., 2001;Shapiro et al., 1991;Heerey et al., 2005). However, playful teasing in infancy has not received much attention in modern developmental psychology-possibly influenced by the broadly negative value accorded to the word teasing in Anglo-American cultures combined with an implicit belief in childhood innocence. Studies of playful teasing assume that it emerges in middle childhood, requiring more complex socio-cognitive skills than does more malicious teasing and becoming more playful and prosocial with age (Warm, 1997;Keltner et al., 2001;Shapiro et al., 1991;Heerey et al., 2005).
The few studies of playful teasing in infancy generally focus on parents teasing infants (Nakano & Kinaya, 1993;Reissland, Shepherd & Herrera, 2005;Labrell, 1993; see also Göncü, Mistry & Mosier, 2000 for reports of variability across communities) rather than on teasing by infants, a behaviour which is harder to observe. And yet, the few reports that exist suggest that teasing by infants can begin with a range of other cheeky and mischievous behaviours around 9 months of age (Reddy, 1991(Reddy, , 2007(Reddy, , 2008Stern, 1985). These studies report early teasing to be largely positive and playful, but since it is usually the parents who are recipients of the teasing their reactions of indulgent and appreciative amusement must at least partially constitute the positiveness of the interaction. The examples reported of teasing which persisted in the face of negative reactions from the recipient were usually directed by infants to siblings and involved snatching their dummies or favourite toys or changing TV channels (Reddy, 2008). In brief, the reports of teasing emerge from around 8 months of age, become clearer and more pronounced by 11 months, and continue through the second year although varying in content and variety.
The ways in which infants tease their parents and siblings vary enormously and involve anything which is of significance in the family's interactions. Typically, three broad kinds of playful teasing have been identified: teasing with offer and withdrawal of objects, teasing with provocative non-compliance and teasing with disruption (Reddy, Williams & Vaughan, 2002;Reddy, 2008;Eckert, Winkler & Cartmill, 2020). However, these categories do not do justice to the variety of ways in which infants can and do tease-a variety which shows the flexibility and, paradoxically, the ubiquity of playfulness and mischief.
From the data in two longitudinal studies of interpersonal play (Reddy, 1991(Reddy, , 1998Reddy et al., 2002) we report clear examples of several types of playful teasing which involve pretence. The data consisted of parental observations of incidents reported during interviews or recorded on dictaphones between interview sessions. As playful teasing often occurs during intimate interactions and is not easily elicited on demand it is not possible to elicit sufficient numbers of video recorded incidents of playful teasing for video to constitute the primary data. Parents were therefore trained to observe and report details of incidents including infant facial expressions, gaze and vocalisations as well as preceding events and responses and the history of that specific type of incident. Probe questions were used in interviews to explore details of the reports. Analyses focused on the reported details of each incident and not on sum-1 3 mary judgements of intention. As described below in each category, details of gaze and expression as well as evidence of prior skill at the serious version of the act were necessary for establishing playful pretend teasing and for ensuring that the acts were not simply errors, performed with a different serious intent or accidental.
We omitted playful teasing with disruption (which we have argued does not need to involve pretence) and grouped the examples into four broad categories: pretend use of communicative gestures (with three sub-categories), pretend responses to rules and prohibitions (with two sub-categories), pretend use of labels and pretend use of object existence. The different categories we obtained are listed in Table 1. In the text, accompanying each category we present one illustrative example per category. The examples we have chosen are all from around or below 12 months of age (before the period when pretend play with object meanings is usually reported) and in all these examples the reported history and observed details are very clear.

Playful Teasing using…
Action Pretence

Communicative Gestures
Object Offers Pretending to give: Starting to offer an object to someone, then withdrawing it as they reach out Arms Out Pretending to approach: Starting go to someone, then pulling away as they reach Pointing Pretending to want: Requesting a drink then refusing it when offered Prohibitions Almost Non-Compliance Pretending to non-comply: Almost doing a forbidden act but stopping short Almost Compliance Pretending to comply: Nearly (or inadequately) completing the requested act Labels Mis-labelling Pretending to make an error: Calling parent by the wrong name Non-existent Objects Action with nonexistent object Pretending object existence: Picking up imaginary food and putting it in bowl Note: As described in the text, we ascertained that the infants in the examples of all these actions showed contemporaneous evidence of being able to perform the serious versions of the actions and that the actions were accompanied by gaze to recipient and playful or cheeky expressions (to ensure that they were not simply changing their minds or performing the acts in error or with a different intent). The examples in the text were taken from original unpublished interviews, dictaphone records and videos from two studies; these records have been partially used in previous reports in Reddy, 1991Reddy, , 2007Reddy, and 2008. Use of communicative gestures Playful pretend teasing with three kinds of communicative gestures are noted below: object offers, arms out, and pointing. To establish that the gesture was being playfully 'mis-used' it was first established in each case that the infant could and reliably did engage in the conventional version of the gesture at other times, and that the infant's facial expressions prior to and after the act indicated that it was playful rather than a change of mind.
Object offers Holding an object out towards someone, acting as if you are offering it to them, and then withdrawing it when they reach out for it might be seen as an almost archetypal act of teasing and common in infancy. It has also been observed in chimpanzees (Koehler, 1927;Call & Tomasello, 2007), bonobos (Krupenye et al., 2018) and an orangutan (Cartmill & Byrne, 2010). The human infant's facial expression while holding out the object is often described as cheeky or with a watchful half-smile. In the video-recorded example below the infant was 9 months old and had been properly offering and giving objects for about two weeks.
S is seated in a high chair, her father on a chair beside her at the table. The tease follows several successful instances of the offer and release of a small biscuit by both father and S, each exchange accompanied by smiles and "Ta". S stretches out her arm once again to her father holding out the little biscuit, her eyes on his face, now watchful with a slight smile on her face. He obediently (but perhaps now wanting to get on with his dessert) stretches his arm out for it again. As his hand starts to approach she pulls hers back, smiling more and wrinkling her nose. He is surprised and laughs, saying "Give me, gimme, gimme!" and reaches further forward for it. She pulls it back further, smiling, her gaze on his face, then turns her body away from him briefly in the high chair, then turns back to him. He withdraws his arm, but is still looking at her. Her eyes are still on his face. She stretches her arm towards him again offering the biscuit, watching his face with a half-smile and as he starts to reach out in response, she quickly whips it back and turns briefly away again.Described from video of 9 month-old S with her father (Reddy, 1991).
Arms out This is a less common version of offering and withdrawing an object, where the infant, either spontaneously or in response to a request to go to someone, holds her arms out as if to go to the other and then pulls back laughing as they respond (a reverse situation to the fausse sortie tease reported by Groos, 1896Groos, /1976 where the child first playfully refuses and then goes to the other). The example below was obtained during interview with the mother of an 11 month-old infant.
" . she's been crawling more now… … And it's usually either in the morning when we're both here or like when I'm just going off to work. And she's got this habit now of like, say I'm holding her and my husband come up to get her and she'll pretend to go to him and then she'll back off. And that's with anyone, like even if my Mum comes round and I say 'go to nanny', she'll like put her hands out to go and then she'll back off…" "She will laugh, she thinks that's really funny, because they go 'oh, um' and she'd keep doing it…" been happening-1 3 within the last month… "She's looking at the person she's going to go to, 'cos they're speaking to her 'come on then' like that and they put their arms out to get her and she'll go to turn towards them and then like back off and laugh over my shoulder.Maternal report in Interview of 11 month-old D.
Pointing The classic pointing gesture involves outstretched arm with extended index finger and the other fingers curled under the palm (although many variations exist, see Kettner, 2021) and can be used to serve different functions: requesting, informing or questioning. The playful misuse of pointing can potentially involve any of its typical functions. The example below shows an 11 month-old infant using the recently developed gesture of proto-imperative pointing by repeatedly pretending to request a drink and then smilingly refusing it.

Teasing involving rules and prohibitions
Teasing with provocative non-compliance is very common in human infancy, but not clearly reported in other primates. Although teasing is also sometimes reported by parents in relation to positive directives such as 'Wave to aunty', 'Show how you clap your hands' etc., it is easier to establish playful teasing in relation to prohibitions involving a 'No!' or a 'Don't do X'. To establish that the act was indeed a playful tease rather than a straightforward desire for the forbidden thing, it was first established that the infant could and did, at least on occasion, desist when told 'No' from some prohibited behaviours, and from the specific behaviour in the instance reported. The infant's gaze, facial expressions and subsequent behaviour were explored for establishing whether the infant's primary interest was in the forbidden act or in the other's reaction to the act. To constitute pretence rather than just provocation, the action needed to be incomplete in some way-begun but not completed or suggesting non-compliance-acting 'as if' the infant was going to non-comply rather than actually or already doing so.
Almost non-compliance Burk (1897) refers to 'almost' or 'pretend' performance of the forbidden as evidence of a primitive antagonism to authority, but nonetheless reporting the roguish or twinkling looks and laughter accompanying such acts in preschool children. One such infant example is reported below where the infant (previ-1 3 ously known to comply to prohibitions about touching hot objects) flicks her finger towards a hot teacup, intently watching her mother's face as she does so.

M:"She teases with almost doing something she shouldn't. If you've said "don't touch the cup it's hot", she'll tap it and take her hand away and look at you as
if to say I'll see how hot it is, you know. She just flicks it, you know, taps it-a kind of cheekiness". F: "She doesn't touch the radiator (anymore)" M: "She just doesn't really touch it. …. She gets cocky now, as I say, if you have got a cup of something in your hand and you say it is hot and she will go (action of flicking finger towards cup while looking at the mother)…. " "As I say it is not a smiling face but it is not a miserable face, it is that sort of half and half, but you can see there is an evil look in her eye that she wants to do something and she is going to just push you to see quite how far".Parents' Interview report of 11 month-old V.
Almost compliance Interestingly, playful teasing in relation to rules or prohibitions can also be accomplished through partial or almost compliance-where you almost do what is asked of you, but not quite, an acting as if complying. One such example is described below where in response to a familiar prohibition about not screaming, the 13 month-old infant looks at the parent and gives a voiceless scream: "during lunch time. One of the things Jonathan is doing a great deal now is screaming and he is doing far more than he has ever done it before. On this lunch time he gave a huge scream, really piercing and I turned round and said 'Don't scream' and he just opened his mouth and gave a sort of voiceless scream which I have to admit I found rather amusing and I did laugh …. At the time he was simply waiting for me to give him some food. His mood was good and it was just that this screaming does occur so often. He is continuing to scream a lot which, I mean I know babies of this age do a lot. I have only once since had him um… doing a scream and me saying don't scream and he has actually given a sort of voiceless scream again"Mother's Dictaphone report of J at 13 months, 9 days: Teasing involving labelling Chukovsky reported a classic tease with a deliberate error-of his 16-month-old daughter laughingly saying "Doggie Miaow" (although she knew well that it was cats which went "Miaow"). The mislabelling can also be done with simple object names as in the case reported below of an 11 month-old cheekily (and repeatedly) referring to her mother as Daddy.
AL, had been confidently and correctly naming herself as 'Baby' and had already been correctly naming her mother as Mummy and her father as Daddy. One day she pointed to her arm, said 'Baby', then suddenly pointing to her mother said 'Daddy'. The mother's puzzled correction led only to an insistence on the 'error' until her 'cheeky look' with 'her head on her side' gave the game away.Mother's interview report of AL at 11 months: Teasing involving object existence Finally, there are occasions when a playful tease can involve a pretence using object meanings-as in the more typical approaches to pretend play. Described below, and also reported for instance by Palacios and Rodriguez (2015) is an example of a 12 month-old infant smilingly picking up pretend food and putting the non-existent food into a bowl.
"…during tea-time. Giving J his tea I had a friend with me, K, who was also with me when I was feeding him and whilst giving him his food out of a bowl with the spoon I was feeding him, he was also holding a spoon at the same time which we often do if I need to distract him a bit or interest him to eat a bit more. He then pretended to push his finger into the bowl, sorry, he pretended to put food into the bowl… that wasn't there… he just picked something up from his highchair and pretended to put some food into the bowl and just grinned looking both at me and K in a rather cheeky sort of way, but obviously aware that he didn't actually put anything into the bowl. … We laughed, so this continued several times during tea-time."Mother's Dictaphone report of J at 12 months, 18 days: We have focused on this selection of examples for the purpose of discussing the pretence involved. In doing so, however, we have ignored many other examples of playful teasing in which the tease seems to precede the grasp of the conventional meaning or the reliable ability to perform the conventional version of the act and thus makes the link to a pretence less clear. However, these earlier episodes, also very cheeky and playful in the performance of similar actions, seem to be setting up engagements which could later involve clear pretence. In the case of object offer and withdrawal there were some examples of infants cheekily holding out objects and withdrawing them before they ever showed evidence of being able to let go of a proffered object. For instance, N, at 7 months and 20 days had never actually offered and released objects before. She held out a 'weeble' (a little toy man) and, her eyes on her mother's face and her expression watchful and cheeky, and as the mother reached out for it, dropped it in the gap between her own and her mother's hand and giggled. She repeated this offer and dropping one more time, but the attempt did not develop as a game. Since she had not yet started giving objects, the cheeky dropping of the toy was not a clear tease nor quite a pretend offer. Similarly, there were many examples of cheeky non-compliance to prohibited actions at 8 and 9 months; although the infants showed awareness of the prohibition they had never actually complied with it, and seemed as motivated by the prohibited action itself as by the adult's reactions. For instance, at 10 months and 2 days, C had never complied with a "No!" and her mother tended to just distract her. Her mother reported one particular favourite forbidden action of touching the video player/ recorder in a teasing way: "…. cheeky, very much so, she knows she's not supposed to do it… she'll keep looking at you and she's got her hand there in amongst it and she's sort of pushing her hand in the flap and she's looking at you, you know, 'oh look, look, I've got my hand there' … she knows she's not supposed to do it and she's saying well, look, I'm doing it". The non-compliance was playful and provocative but it was explicit-not pretended. Although not counted as playful pretend teasing, these examples nonetheless suggest 1 3 that the practice of cheekiness in the performance of interpersonal actions is wellestablished before we can clearly identify explicit pretending as in the examples in Table 1.

Implications for the emergence of pretending
We ask two questions in this section. First, how do these examples of playful pretend teasing compare with pretend play with objects? And second, given that playful teasing pretend occurs a few months earlier than does pretend play with objects, what does this mean for understanding the emergence of pretence itself?
In the examples of playful pretend teasing in Table 1, the infants have already grasped the conventional 'meaning' of the acts which they are using to tease the other. As evident from their normal usage of the acts at other times-of offering objects, holding arms out, pointing, complying or not complying and labelling-they have grasped what each act 'signifies' in terms of the response from the other, the next follow through step from the infant, and so on. The normal versions of these acts have recently entered the infants' repertoire of actions. When the communicative gestures or the labels are 'mis-used' in playful teasing, their meaning is distorted from the conventional usage. Instead of an object offer meaning something like 'I am holding this out to give it to you', it now seems to mean something like 'I am holding this as if I am going to give it to you'. In the same way as Jacqueline was 'pretending to sleep' by using the 'putting head on pillow' act when not actually intending to sleep and not at bedtime (Piaget, 1962), the infant is using the communicative gesture in a different action context-pretending to give; in going beyond the conventional usage and violating the meanings, this use of the act reveals symbolic qualities (Palacios & Rodriguez, 2015). In the case of the playful teasing using 'almost non-compliant' or 'almost compliant' acts, the minimised or modified form of the acts appears to be a reference to the 'normal' version of the infant's own non-compliance or compliance at other times. The 'almost non-compliance' doesn't mean 'I am going to do this even though I know you don't like it' but rather something like 'I am acting 'as if' I am going to do this because I know you don't like it' or 'because I want to see what you will do' and so on. If one were to use the language of de-coupling one could see the teasing versions of all these acts as de-coupled from their conventional significance. The infant at different times and sometimes in the same episode, moves between the conventional version of the act and the playful violation. These violations are motivated by a desire to provoke emotional reactions in others-reactions which vary between recipients and which invariably change the meaning of the playful pretend tease, making it a regular game, a battle or terminating it entirely, as time goes on.
However, the infant's grasp of what the act signifies is very contextual: it appears to be restricted to the act when performed by the infant, but not necessarily when receiving it. Some infants were reported to get 'quite cross' on occasion if the parent offered and withdrew objects, or just to stare uncomprehendingly if the parent 'pretended' in the 'same way' as the infant, consistent with findings about infant awareness of others' playful intentions emerging in the second year (Nakano & Kanaya, 1993). For understanding the playful violation of the conventional meaning of actions 1 3 when done by others, infants seem to need more of a build-up than when they do it themselves. It is unclear though, whether in the case of pretence with object meanings, toddlers are as comfortable with their own pretence as they are with others'. It is very likely that the same disparity of understanding pertains to object pretend play too (Lillard, 2007).
In both cases, that is in relation to pretend play with objects as well as pretence in playful teasing, the content of the pretence in infancy seems to concern recently mastered conventions or recent foci of interest. One function of pretend play has been argued to be a making sense of one's experiences (Göncü & Perone, 2005), pretending enabling one to explore and test the boundaries or substance of something that is of current focus. Many of the teasing examples reported here involve newly developed understandings which are played with-the newly mastered gesture of offering and giving objects, the newly developing understanding of a specific prohibitionand are pushed playfully and provocatively to their limits. As with adult teasing, the same (successful) types of teases may be repeated at older ages, sometimes with more complex setting-up and contexts, and sometimes develop into games where the tease is no longer potent and may no longer function as a tease or provocation. The origin of teases seems to lie, however, in the early period around the time that the conventional meaning that it is playing with is first engaged with. The pre-verbal negotiation of 'intersubjective truth' and frames occurs in early social engagements well before explicit language-based skills develop in the 2nd year of life (Karniol, 2016).
The proposal that playfighting in other animals could be considered an analogue (and phylogenetically early form) of pretend play (Lillard, 2017) is worth comparing to the phenomenon of playful teasing. Lillard offers three criteria for comparing playfighting and pretend play which could also be useful in relation to understanding playful teasing. Playfighting behaviours, she argues are, firstly, "issued and understood 'as if' they were real fighting behaviors" with one reality being imposed over another but without any confusion between the 'as if' and the 'real' worlds; secondly, the play intention is marked by recognisable play rituals; and thirdly, both 'fighters' share the knowledge of the play intention (Lillard, 2017).
Playful pretend teasing appears to go beyond the basic requirements of these criteria, suggesting that by the end of the first year it is already more complex in relation to pretend play than is play fighting; but it also reveals important differences. As in play fighting, in the playful pretend teasing incidents the acts are issued 'as if' they are real; the playful violation of the conventional meaning is not done in error or some other confusion with the reality; and once the playful violations are noticed by the recipient (whether responded to positively or negatively) the teasing intention becomes shared knowledge. In contrast to playfighting, however, the 'markers' of playfulness in these examples-cheeky looks and half-smiles which can reveal the playful intention-are not quite the ritual markers of playfighting. They are different from the 'play bows' of dogs or the neck-nuzzling in rats-which are "communicated in specific, ritualised ways so participants know clearly that they are partaking in the as-if rather than the real world" (Lillard, 2017, p. 828). In playful pretend teasing, the acts appear intended to fool the other into reacting as if the act was real, rather than as markers of 'this is play' (Keltner et al., 2001); and in some cases-e.g., with the communicative gestures-the adult is indeed fooled, at which point the cheeky 1 3 expression gives the game away. The minimised acts in 'almost non-compliance' or 'almost compliance', however could be seen as ritualised markers in some way (see also Bates et al., 1978 for discussion of the emergence of ritual actions in infants). Flicking the teacup or moving the hand almost to the plant but not quite-accompanied by watchful looks to the parent, or sometimes using attention-getters to obtain the parent's attention before doing so, are different from actual non-compliance when the infant simply does the prohibited act. Again, in these cases the infant may be using the 'almost' non-compliant act as a 'nip' rather than a bite (to use Gregory Bateson's example), or 'de-coupling' them. Similarly, the silent scream in response to the prohibition in Table 1 could be seen as a reference to the actual scream and to the compliance. Lillard's discussion of playfighting as analogue and early form of pretence brings one crucial difference from object pretend play into focus-like playful teasing it too depends on another 'person' being a recipient.
Pretend play with objects, as widely acknowledged, emerges at the earliest "between 12 and 24 months of age, and from 2 to 3 years, it becomes well-established" (Lillard, 2007, p. 168). Playful pretend teasing, however, emerges from around 9 months of age, becoming clearer by around 11 or 12 months. What does this mean for our understanding of the emergence of pretence in human infants? It suggests that the roots of pretend play lie much earlier, in the second half of the first year, than conventionally believed, and suggests a more continuous emergence of the symbol embedded within engagements with people. The argument for a continuous emergence of the symbol with early acts being partially symbolic is not a new one; nor is the suggestion that symbols emerge in the service of communicative functions (Bates, 1977;Werner & Kaplan, 1963;Trevarthen & Logotheti, 1987;Palacios & Rodriguez, 2015). However, our argument confirms and extends recent views that put people at the centre of the story about how pretending becomes possible (Haight & Miller, 1992). Not only do engagements with people set up the communicative abilities and the conventions and usages for the capacity for pretence, but we suggest that it is the emotions involved in playing with people which enable the emergence of pretending 3 .
There are different reasons for bringing relations with persons into the centre of the emergence of pretending even when talking of pretend play with objects in the second year. In a broad sense, the mutuality of relations with people and relations with objects is clear in all object usage (Costall, 2013;Palacios & Rodriguez, 2015;Williams, Costall & Reddy, 2018). Contrary to previous assumptions in developmental psychology (Fein, 1981), studies of pretend play in urban western communities (but not necessarily in other cultures) show that relations with people are crucial even when talking specifically about pretend play with objects: Early pretend play with objects in infancy is usually initiated by mothers (Haight & Miller, 1992); joint pretending episodes at 2 or 2-and-a-half years of age are longer than solo pretending episodes (Haight & Miller, 1992;Dunn & Wooding, 1977;Slade, 1987); the pres-1 3 ence of older siblings enhances the rate of pretend play (Dunn, 1988); and severely deprived early interpersonal connections are linked with paucity in pretend play (Lillard, 2017). More importantly, as recently argued by Palacios and Rodriguez, "the first symbols are not rooted in a literal evident reality but in a world of shared rules of uses about the material world" (Palacios & Rodriguez, 2015, p. 23). Rather, they argue, pretend play with objects emerges in infants because of adult usage of objects and adult involvement in engaging with infants' uses.
The role of the interpersonal in relation to playful pretend teasing is even more significant, with one necessary feature marking it out from all other pretend play. Teasing is done to as well as with people. It cannot be performed alone. In describing Jacqueline pretending to sleep, Piaget may well have forgotten to observe the presence and role of other people in the pretence. But it is also possible that Jacqueline actually did play with the pretend act on her own, exploring the act itself. In playful pretend teasing, however, the motive for the 'mis-use' of communicative gestures is precisely to draw a reaction from the other. It is the other's reaction to the violated usage that attracts the infant, rather than an interest in the violation itself. Although infants can explore actions on their own (e.g., infants touching a prohibited object and shaking their heads or saying 'No' to themselves) in playful teasing the infant's interest is specifically in eliciting the other's reaction. The eliciting of others' reactions is very likely rooted in much earlier playful engagements involving surprise such as peekaboo (Ratner & Bruner, 1978). Infants are not only drawn into the practice of such socio-cultural games in the first few months of life , but very soon begin to be active participants in the creation of surprise (Nomikou, Leonardi, Radkowska, Racaszek-Leonardi & Rohlfing, 2017;Rossmanith, 2018). Playful teasing of all kinds may be an exploration of emotional stances-an exploration which would not be possible if people didn't react. In conventional engagements such as giving and taking of objects, prohibiting the pulling of hair or touching of a video recorder, requesting a wave or other gesture, the adult's role (of taker or offerer or prohibiter or requester) and their emotional stance (of approving or encouraging or prohibiting and so on) is evident and evidently different from the infant's. Similarly, the potential for emotional reactions to different actions must also be present and evident. If people didn't take when you offered, or if they weren't interested once you withdrew it, or if people didn't say 'no' and follow through one way or another when you didn't comply with the no-teasing such as we see it in typically developing human infants could not emerge. It is only because other people allow the existence of violation and provocation, because their attitudes themselves are interesting and because they have emotional reactions to being teased that interpersonal gestures and agreements start to be violated and played with.
Playful engagements with people, especially familiar people who enjoy you and offer a range of complex intersubjective engagements Trevarthen & Hubley, 1978), create the crucial ground for the emergence of pretence. Infants' early familiarity with people's attitudes and emotional reactions towards the infant makes them easier to play with and easier to test and violate. Symbols may well arise from a shared world of rules of usage of objects (Palacios & Rodriguez, 2015) or from the infant's perception of the others' orientations towards shared situations affording the possibilities for different stances in relation to object meanings (Hob-1 3 son, 1990). Even more important for our understanding of pretend play, however, is the potential of early playful engagements between infant and adult for creating for the infant the very possibility of playful violation of meanings in pretending. Playful pretend teasing maybe a first step for the infant in learning how to pretend.