The APM proposes that the α-dimensions derive each from one corresponding caregiving feature (Table 2), which we term α-features. To account for such a causal link, we suggest the following – more specific – definition of avoidance and ambivalence starting from a precise definition of the corresponding α-features – that we refer to as sensitivity and responsiveness respectively:

These definitions entail a two-channel-hypothesis – i.e., that two main (relatively) independent communication channels, one emotional and the other physical, are first relevant in the attachment relationship – with a one-to-one causal link between caregiving features and attachment dimensions:

The insensitive caregiver does not engage emotionally and encourages the child to deactivate attachment. As a result, the child tends to avoid interactions based on attachment. The unresponsive caregiver does not engage physically and encourages the child to hyper-activate attachment. As a result, the child tends to be worried about the caregiver’s availability (with evident emotional display).

Importantly, these definitions provide a clear correspondence between caregiving features and attachment dimensions. Insensitivity and avoidance match with each other by virtue of their emotional nature. The avoidant avoids the intolerable experience of facing a cold caregiver by deactivating attachment and focusing on other activities (Ainsworth et al., 1978; George et al., 1985; Hesse, 2008; Mikulincer and Shaver, 2016). A child will typically explore or play. When the insensitive caregiver is present, they do not connect emotionally and, as a result, asking comfort from them makes no sense. Unresponsiveness and ambivalence match with each other by virtue of their physical nature. The ambivalent hyper-activates attachment and worries whether their caregiver will be physically there for them (Ainsworth et al., 1978; George et al., 1985; Hesse, 2008; Mikulincer and Shaver, 2016). The unresponsive caregiver is unpredictably available but connects emotionally when present and, as a result, an emotional signal demanding their presence makes perfect sense. Interestingly, an insensitive-unresponsive caregiver will correspond to an avoidant-ambivalent child, but the two dimensions cannot be expressed simultaneously (since deactivation and hyper-activation are incompatible).

Hence, the provided definitions perfectly illustrate the basic characteristics of the attachment relationship. Their adequacy is further supported by the following four arguments.

(1) Statistical-independence argument. As discussed above, avoidance and ambivalence as defined in the literature are generally considered mutually (relatively) independent. We define avoidance and ambivalence more specifically – referring to the caregiver’s sensitivity and responsiveness – but still maintain that they are (relatively) independent. In fact, if they are, there must be two mutually (relatively) independent caregiving features, each of which induces one of them, and they also need to be identified.

The definitions given above support independence both between insensitivity and unresponsiveness and between avoidance and ambivalence. Indeed:

The given definitions meet the requirement of reciprocal (relative) independence. By contrast, the literature has not focused on the mutual independence of caregiving features. Distinguishing between a purely emotional channel and a purely physical one allows us to ensure such independence.

(2) Developmental argument. We consider two points. (A) In order to develop adequately, for the child, both emotional connection and physical care are essential (Bowlby, 1969/1982, 1973, 1980; Stern, 1985; Schore, 1994; Sroufe, 1995; Leerkes and Wong, 2012; Marvin et al., 2016; Feldman, 2017). (B) As mentioned above, some caregiving features have been suggested as inducing avoidance and ambivalence. In particular, Bernier et al. (2014) found that sensitivity and autonomy support as defined by Whipple et al. (2011, p. 397), although not reciprocally independent, fully explain the generation of avoidance and ambivalence. This entails that it is possible to find two caregiving features on which avoidance and ambivalence entirely depend.

These data suggest that we should seek the cause of avoidance and ambivalence considering (A) the emotional and physical aspects of development and (B) two caregiving features. This is, indeed, the content of our proposal.

(3) Evolutionary argument. We have reason to believe that in the Environment of Evolutionary Adaptedness (EEA), the context of human evolution (Bowlby, 1969/1982), avoidance and ambivalence have been transmitted through two mutually independent channels – one emotional and the other physical. More precisely, according to an evolutionary argument, avoidance derives from the unwillingness to invest and ambivalence from the inability to invest in the offspring (Chisholm, 1996; Chisholm and Sieff, 2014).

On the one hand, unwillingness is an emotional feature that expresses intentional rejection. Evolutionarily, this is due to a harsh environment that makes the caregiver opt for investing in mating as opposed to parenting in order to maximize their reproductive success. In this condition, the child finds the best fit ignoring their caregiver and boosting their autonomy (avoidance).

On the other, inability is a physical feature that expresses the impossibility to be physically there. Evolutionarily, this is due to an unpredictable environment that occasionally forces the caregiver to attend to essential survival activities, thereby abdicating their role. In this condition, the child finds the best fit amplifying their need signals (ambivalence).

The evolutionary argument is supported by developmental evidence. Indeed, caregiver willingness and ability seem to be crucial information for young children as proved by the ability of 9-month-old infants to distinguish between the unwillingness and inability of an adult to hand them a toy (Behne et al., 2005).

(4) Neuroscientific argument. We have defined: (A) Insensitivity and avoidance as socio-emotional properties. As such, they primarily rely on gaze direction and facial expressions (relevant to emotional connection). (B) Unresponsiveness and ambivalence as socio-physical properties. As such, they primarily rely on reciprocal position (relevant to physical protection and care).

Two quite independent brain networks can be identified as underpinning these emotional and physical aspects. (A) On the one hand, the superior temporal sulcus, which is essential to detect gaze direction (Pelphrey et al., 2005; Hoehl et al., 2009; Itier and Batty, 2009; Carlin and Calder, 2013), is connected to the amygdala, which is key to reading facial emotions (Loughead et al., 2008; Whalen et al., 2013; Gothard, 2014; Wang et al., 2017). (B) On the other, the precuneus region is essential to deem the reciprocal position (Peer et al., 2015).

These neuroscientific data are consistent with two (relatively) independent channels, one emotional and the other physical, that underpin insensitivity/avoidance and unresponsiveness/ambivalence respectively.

Concluding this section, we want to stress that the focus on these two basic channels is proposed to help further understand and differentiate the complex nature of the two fundamental dimensions of avoidance and ambivalence and not to reduce it to such channels. In other words, the two channels are suggested to catch a fundamental distinction between the two dimensions and not to describe them exhaustively.

Overall, the above arguments converge to strongly support the two-channel-hypothesis. With these definitions, three attachment dimensions and three corresponding caregiving features are revealed to be clearly causally related: (1) a frightening caregiver induces disorganization, (2) an insensitive caregiver induces avoidance, and (3) an unresponsive caregiver induces ambivalence. Accordingly, each α-dimension can be thought of as corresponding to a core belief: “My caregiver is frightening,” “My caregiver is insensitive,” and “My caregiver is unresponsive” respectively (Table 3). Therefore, the APM indicates how attachment – and the related aspects of personality – are transmitted from a generation to the next, thereby solving the intergenerational transmission problem (P1). However, three dimensions are still insufficient to account for the relationship between attachment and psychopathology – we suggest these dimensions to be seven.

We define phobicity as the subjective measure of how much the caregiver is limiting (in terms of the child’s exploration). The child needs to get their tendencies to attach and explore regulated by the caregiver, and the phobic child experiences a limitation of their exploration. In other words, the caregiver somehow forces the child to be closer to them than the child feels necessary. For example, the caregiver may be hyper-protective and tend to keep the child under their strict control. In this case, the child will tend to feel restricted and, when confronted with the task of autonomously exploring the environment, particularly in danger.

As a result, we propose, the preschool child gets imprinted an implicit knowledge that can be expressed by the core belief “I am in danger if my caregiver is not with me” and “My caregiver won’t let me go” (Table 3). This knowledge characterizes phobicity making the phobic particularly sensitive to the balance attachment-exploration and inclined to suffer from (1) separation anxiety when too far from the caregiver and (2) a sense of constriction when too close. In practice, phobicity is revealed by such sensitivity.

Our definition and proposal are supported by the following evidence.

We define depressivity as the subjective measure of how much the caregiver is emotionally unreachable. The child needs to be able to reach the caregiver for emotional care, and the depressive child experiences a failure of their attempts to do so. In other words, the caregiver is, for some reason, emotionally unavailable to the child when the child tries to reach them. For example, the caregiver may be usually away from home. In this case, the child will tend to see their desire for emotional care frustrated and, therefore, to feel hopeless, defeated (Seligman, 1975; Guidano and Liotti, 1983; Guidano, 1987, 1991, 2007).

As a result, we propose, the preschool child gets imprinted an implicit knowledge that can be expressed by the core belief “I won’t be able to reach my caregiver” (Table 3). This knowledge characterizes depressivity making the depressive particularly sensitive to the loss of the caregiver and inclined to suffer from (1) a sense of defeat and (2) depression when a loss is perceived. In practice, depressivity is revealed by such sensitivity.

Our definition and proposal are supported by the following evidence.

In a difficult EEA, the child that gets imprinted to give up attempting to reach an emotionally unavailable caregiver enhances their chances to survive by becoming self-reliant (Bowlby, 1980). A tendency to self-reliance starts to be feasible when the child begins to develop some independence (β-period, 2–6 years). Later, the depressive adult will have the advantage to be used to relying on themselves, thereby improving their survival and reproductive chances in a harsh competitive environment.

We define somaticity as the subjective measure of how much the caregiver is defining (in terms of the child’s internal states). The child needs to learn to recognize their internal states – primarily, the most somatic ones: sensations and emotions – from the caregiver, and the somatic child experiences, instead, an external definition. In other words, the caregiver somehow forces the child to adopt an internal state that does not match the child’s actual one. For example, the child might be suggested they feel pain when they do not (like in the case of Brenda, Guidano, 1987). In this case, the child will tend to feel confused about their own state and intruded upon by the caregiver. Somaticity is, therefore, an anomaly of the fundamental caregiving regulation of the child’s internal states, where the caregiver’s state does not correspond to the child’s state. This way, the child does not learn how to recognize their own internal states and becomes uncertain of them.

As a result, we propose, the preschool child gets imprinted an implicit knowledge that can be expressed by the core belief “I need my caregiver to tell me about myself” and “My caregiver will intrude on me” (Table 3). This knowledge characterizes somaticity making the somatic particularly sensitive to the definition of their own internal states and inclined to suffer from (1) uncertainty when not sufficiently defined and (2) a sense of intrusion when being excessively defined by the caregiver. In practice, somaticity is revealed by such sensitivity.

Our definition and proposal are supported by the following evidence.

We define obsessivity as the subjective measure of how much the caregiver is judgmental. The child needs to learn ethics – namely, what is considered right or wrong – from the caregiver, and the obsessive child experiences the imposition of the caregiver’s code of conduct and being significantly blamed for not abiding by it. In other words, the caregiver enforces a strict and arbitrary set of rules by systematically blaming the child for disobedience. The blame is always justified by claiming to cause terrible harm to someone and conveys the implicit message of the caregiver’s rejection of the child. It can assume different forms, such as verbal scold or physical punishment. For example, the child might be severely reproached for wearing wrinkled clothes (like in the case of Alison, Guidano, 1987). In this case, the child will tend to feel anguished by the possibility of misbehaving – even involuntarily – and to focus on acting correctly.

As a result, we propose, the preschool child gets imprinted an implicit knowledge that can be expressed by the core belief “I am wicked” (Table 3). This knowledge characterizes obsessivity making the obsessive particularly sensitive to ethical matters and inclined to suffer from (1) obsessions focused on their responsibility for causing harm and (2) compulsions to get rid of such intrusive ideas. In practice, obsessivity is revealed by such sensitivity.

Our definition and proposal are supported by the following evidence.

The above arguments strongly suggest that: (A) The caregiver has four fundamental tasks: (1) regulating the child’s balance between attachment and exploration; (2) being emotionally reachable for the child; (3) regulating the child’s internal states and supporting the child’s own definition of them; (4) providing an ethical reference to the child. (B) These tasks become relevant during the preschool years. (C) The child needs to adapt to their caregiver’s particular way of accomplishing these tasks. (D) An optimal solution to this adaptation problem is to detect the caregiver’s attitude with respect to each of these tasks and rapidly and stably acquire such vital information – in other words, to imprint some core beliefs. Given their origin in the attachment relationship, these beliefs correspond to attachment dimensions and, given their fundamental socio-psychological role, they constitute the core of personality. (E) Particular kinds of accomplishments of these caregiving tasks induce the acquisition of core beliefs that can more easily lead to later dysfunctional states and behaviors – i.e., to psychopathology. (F) In particular, using the definitions that we propose: (1) A limiting caregiver induces phobicity, (2) an unreachable caregiver induces depressivity, (3) a defining caregiver induces somaticity, and (4) a judgmental caregiver induces obsessivity. Four caregiving features result to be causally related to four attachment dimensions.

Attachment theory could not find a specific connection between attachment and most psychopathologies because it investigated only the α-dimensions. In contrast, by considering the β-dimensions, the APM can indicate the specific causal relationship between attachment – and, therefore, personality – and the most common mental disorders (Table 5), thereby solving the problem (P1) of fully relating attachment to psychopathology.

Imprinting and sensitive periods have been documented in various species of birds and mammals (Lorenz, 1937, 1981; Harlow, 1959; Hess, 1959; Shipley, 1963; Salzen, 1967; Table 6).

Goslings, for example, attach to their mother during a well-defined brief sensitive period shortly after hatching. Lorenz demonstrated that, when they catch some cues from the environment at the right time, they follow whoever matches the expected features. Famously, they attached to Lorenz himself. Through this – clearly evolutionarily preordained – process, the animal fixes in their mind the identity of their caregiver. Reasonably, given our psycho-social complexity, we-humans might have evolved a more elaborated version of imprinting compared to other animals. Indeed, we argue that we are not only preordained to fix in our mind the identity of specific caregivers, but also the most evolutionarily relevant attachment characteristics – the seven attachment dimensions.

Bowlby (1969/1982) suggested the existence of imprinting and sensitive periods in humans noticing the striking similarities between our attachment and that of other animals, such as the geese studied by Lorenz (1937) or the rhesus monkeys studied by Harlow (1959). Since then, these concepts have been generally maintained by developmental-attachment research (Marvin et al., 2016), but the multiple intervening factors that can contribute to change seem to have caught much more attention than them. However, the α-period is unequivocally confirmed as sensitive for attachment by studies concerning early child-institutionalization that prove the long term effect of attachment experiences that occur within the first 2 years of life (Varin et al., 1996; Nelson et al., 2007; Zeanah et al., 2011; Fox, 2014; Troller-Renfree and Fox, 2017; Tables 1, 6). The effect of such experiences can be actually irreversible.

Regarding the β-period (preschool age), we can consider that:

All this is consistent with the imprinting of the four β-dimensions, which provides the child with vital information to adapt to the four human-specific attachment situations that become relevant during the β-period. As discussed above, we suggest that (Table 2):

Clinical research strongly supports the persistence of personality characteristics from childhood to adulthood and the correspondence between caregiving features and acquired attachment characteristics (Berne, 1972; Bowlby, 1973, 1980; Guidano and Liotti, 1983; Guidano, 1987, 1991; Blatt and Levy, 2003; Guidano, 2007; Levy et al., 2015; Yakeley, 2018). The resistance to change is evident in psychopathology. In cognitive psychotherapy, the phenomenon is called ‘neurotic paradox’ (Perdighe and Mancini, 2010), which refers to the apparently inexplicable persistence of mental disorders despite the patients being aware of their self-damaging behaviors. The phenomenon is consistent with dysfunctional knowledge that is imprinted and can be hardly modified. This is exactly the characteristic of the seven attachment dimensions and, in fact, all major psychopathologies can regularly be linked to information acquired in attachment relationships, as shown by clinical accounts that report early attachment information (e.g., Bowlby, 1973, 1980; Guidano and Liotti, 1983; Guidano, 1987, 1991; Oltmanns et al., 2012).

More specifically for the preschool age, as discussed above (cf. “Phobicity and the Limiting Caregiver,” “Depressivity and the Unreachable Caregiver,” Somaticity and the Defining Caregiver, and “Obsessivity and the Judgmental Caregiver”): (1) the β-dimensions are causally related to specific mental disorders; (2) The β-period is the earliest timeframe for the onset of such disorders; (3) These disorders tend to last throughout life; and (4) Parenting has been recognized as one of the major causes of their onset (Hopkins et al., 2013; Whalen et al., 2017). Since imprinting is the mechanism that underpins the long preservation of information, these data strongly suggest the imprinting of the β-dimensions in the preschool years.

Attachment has been designed by evolution as a fundamental adaptation mechanism. Bowlby (1969/1982; 1973; 1980) has first underlined the survival function of attachment for the child and, later, other authors (Chisholm, 1996; Chisholm and Sieff, 2014; Simpson and Belsky, 2016; Szepsenwol and Simpson, 2019; Young et al., 2019) have stressed the reproduction function of it for the adult. Indeed, the ultimate evolutionary goal of the individual is inclusive fitness (Hamilton, 1964), i.e., passing one’s gene to the following generations. These authors argue that the early attachment relationship provides cues of environmental characteristics – such as harshness and unpredictability (Ellis et al., 2009) – and the child unconsciously gathers these cues from the caregiver to set up an adequate strategy for later reproduction. This consideration suggests that attachment information acquired as a child is meant to be durable. Otherwise, it would lose its value for reproductive purposes. This is precisely what an imprinted IWM with its seven attachment dimensions is meant to provide.

In fact, this line of reasoning fits the preschool age. As discussed above (cf. “Phobicity and the Limiting Caregiver,” “Depressivity and the Unreachable Caregiver,” Somaticity and the Defining Caregiver,” and “Obsessivity and the Judgmental Caregiver”), beyond that in childhood, the β-dimensions have an adaptive function in adulthood: (1) Phobicity promotes protection through maintaining physical closeness; (2) Depressivity promotes the ability of being self-reliant; (3) Somaticity promotes familial cohesiveness through uniformity of thought and feeling; (4) Obsessivity promotes conscientiousness and competence through strict moral conduct. All these characteristics provide the adult with an advantage in a harsh and possibly unpredictable environment, thereby justifying imprinting the corresponding information from when it is available – the β-period indeed – to adulthood.

The neural processes underlying attachment during sensitive periods have been identified and studied in different birds and non-human mammals (Knudsen, 2004; Sullivan and Holman, 2010; Landers and Sullivan, 2012; Knudsen, 2013; Nakamori et al., 2013; Roth et al., 2016; Feldman, 2017; Opendak et al., 2017; Opendak and Sullivan, 2019). These studies suggest the possible neural underpinnings of imprinting in the human brain. In particular, Knudsen (2004, 2013) describes how imprinting in sensitive periods corresponds to a shaping of neural networks that is durable but still admits of a later change through a particular interaction with the environment, which is exactly what is here hypothesized for the IWM. In accordance with Bowlby, it has also been established that, during a sensitive period, attachment to a caregiver occurs regardless of the care received, even when the caregiver is abusive (Sullivan and Holman, 2010; Landers and Sullivan, 2012; Opendak et al., 2017; Opendak and Sullivan, 2019), which further confirms the evolutionary programming of the process. Finally, lines of research across neuroscience and clinical psychology (Schore, 1994; Turnbull and Solms, 2003; Schore, 2009) identify the human brain areas involved in imprinting, with a sensitive period within the first 24 months (Tables 1, 6). In accordance with Bowlby, they stress the life-long durability of the phenomenon and its influence on personality.

Although neuroscience only recently has begun to address the attachment relationship directly in humans, some evidence has already been provided for a sensitive β-period. The studies of Rao et al. (2010) and Luby et al. (2016) found that maternal support in the preschool years is significantly correlated with hippocampal growth until adolescence, while maternal support in the following school years does not correlate with it. As a result, these authors explicitly propose that the preschool years are an attachment-related sensitive period. Moreover, Luby et al. (2016) found that preschool maternal support is correlated with emotional regulation in adolescence, with positive support linked to greater regulation ability. These data confirm the hypothesis of the preschool years as a sensitive period for attachment, with caregiving having a role in psychopathology. This is precisely what the APM proposes for the β-period and β-dimensions.

The above five arguments converge to provide compelling support for imprinting as the specific attachment acquisition mechanism. The α-dimensions are imprinted in a sensitive α-period (6–24 months), and the β-dimensions in a sensitive β-period (2–6 years, the preschool age).

Therefore, the APM solves the dilemma of stability (P2) by clarifying that attachment – as it is expected of anything related to personality – is fundamentally stable.

The APM proposes a revision of the current theory that is fully compatible with its achievements but, at the same time, can significantly enhance its explanatory power. In particular, the model suggests the relationship between caregiving features and attachment dimensions (P1), the basic stability of attachment due to imprinting (P2), and its connection to psychopathology (P3). We reach this result by focusing on the knowledge aspect of attachment and relying particularly on clinical research to explain the available data. This allows us to reduce complexity and identify key elements and their causal links. In contrast, standard attachment theory has been usually trying to take into account as many variables as possible relying mostly on statistical studies that cannot identify causal links. As a result, many variables have been considered, but the relationship between them has appeared often unclear.

In particular, we consider an IWM consisting of seven dimensions: 3 α-dimensions – disorganization, avoidance, and ambivalence – first imprinted in an α-period and four β-dimensions – phobicity, depressivity, somaticity, and obsessivity – first imprinted in a β-period (Table 1). We specify the definition of avoidance and ambivalence so that avoidance is transmitted via an emotional channel and ambivalence is transmitted via a physical channel. As a result, the α-dimensions are each induced by a specific caregiving feature. Furthermore, we introduce the four β-dimensions as each also induced by a specific caregiving feature. Consequently, we provide a complete mapping between caregiving and attachment (Table 2). According to the APM, the caregiving features correspond to fundamental tasks, among which the caregiver normally switches according to the situation, thereby inducing the acquisition or the expression of the corresponding attachment dimensions. On the other hand, given that they correspond to different caregiving features, the attachment dimensions are assumed to be independent, at least in terms of their function. However, since the same caregiver is usually responsible for the induction of multiple – if not all – dimensions, some characteristic patterns are to be expected. For example, when somaticity is the predominant dimension, then phobicity is also expected to be high because a defining caregiver will probably somehow limit their child’s exploration (Guidano, 1991). Finally, disorganization and the β-dimensions explain how attachment is linked to psychopathology (Table 5), as further discussed below.

We focus on dimensionality and imprinting as the central aspects of attachment that derive from its evolutionary predefinition. We propose that these seven dimensions specify the full range of adaptation-vital information, and imprinting provides the necessary stability to its acquisition – making it a durable and reliable adaptive base for the child and the future adult. In contrast, standard attachment theory has been exclusively concentrating on the information acquired within the first 2 years of life and on the possibility of its change. As a result, key aspects of attachment could not be identified.

Personality is a complex construct, and numerous models have been proposed from different perspectives. Between them, we focus here on one of the most influential and widely referred to: the Five-Factor Model (FFM) (or Big-Five) (McCrae and Costa, 2003). It consists of five traits – Openness, Conscientiousness, Extraversion, Agreeableness, Neuroticism – that have been statistically extracted from linguistic descriptions of personality characteristics. The model can be applied to healthy subjects but has also been linked to personality disorders (Widiger et al., 2017). An important feature of the FFM compared to the APM is its purely empirical nature with no endorsement of any specific etiological theory of personality. In other words, it is purely descriptive and cannot explain where the dimensions it postulates come from. In contrast, the APM is grounded in attachment theory – a theory of human relationships that finds in early experiences with the caregiver the foundation of personality, thereby offering an evident explanatory advantage. Moreover, in the condition of its applicability, we can expect the APM to have no less descriptive power than the FFM, since the set of α- and β-dimensions covers a broad range of personality features (Bowlby, 1969/1982, 1973, 1980; Guidano and Liotti, 1983; Guidano, 1987, 1991, 2007; Nardi et al., 2010; Liotti and Farina, 2011; Hesse, 2016; Mikulincer and Shaver, 2016). Interestingly, empirical studies have found correlation between the fundamental attachment dimensions of avoidance and ambivalence and the Big Five (Noftle and Shaver, 2006; Fraley et al., 2011). In this regard, the most relevant finding is the negative correlation of avoidance with agreeableness and the positive correlation of ambivalence with neuroticism. These results match the typical avoidant detachment and ambivalent over-emotionality. Given their meaning, we can also expect some correspondence between the β-dimensions and the FFM traits. Somaticity, for example, is expected to be correlated with agreeableness, obsessivity with conscientiousness.

The connection between the APM and the FFM needs to be further investigated and, at the moment, only preliminary hypotheses can be formulated about it. In this respect, a central issue concerns the possibility of personality change throughout life. The APM proposes personality to be resistant to change, but empirical findings underpinned by the FFM indicate that changes are to be expected (Lucas and Donnellan, 2011; Oltmanns et al., 2019). For example, agreeableness and conscientiousness have been found to increase between age 20 and 60, while openness to decrease after 60, with important individual differences (Roberts and Mroczek, 2008). A ‘maturity principle’ has been proposed (Caspi et al., 2005; Lucas and Donnellan, 2011) according to which agreeableness and conscientiousness tend to increase and neuroticism to decrease over young and middle adulthood in relation to the assumption of mature social roles. The end of maturity (Oltmanns et al., 2019) has also been identified, corresponding to the decrease of extraversion, conscientiousness, and openness from late middle age. In particular, personality stability and change over the life course have been connected to identity processes (Roberts and Caspi, 2003). According to this perspective, continuity is related to identity strength, in terms of achievement and certainty, for example. Changes of identity that are characteristic of different stages of life – and related to different social roles (e.g., being independent, having a family, pursuing a career) – would explain corresponding changes of personality (Roberts et al., 2003, 2006). Although the apparent contrast, we can hypothesize that these findings are fully compatible with the general stability endorsed by the APM since the APM and the FFM focus on different constructs. First, the two models seem to measure different social variables related to personality. The APM is based on attachment, which is a social phenomenon limited to close relationships, while the above studies suggest that FFM measures are sensitive to broader social aspects. Second, the APM concerns the knowledge part of personality, whereas the high heritability of the FFM traits (Jang et al., 1996; Power and Pluess, 2015) suggests them to be more sensitive to biological variables and related changes over the lifespan. The indispensable empirical testing of the APM will provide further data to elaborate on this point.

The implications of the APM for psychopathology are even stronger than those for personality.

The APM refers to the aspects of personality related to imprinted attachment data. Therefore, when other variables – such as the biological ones – have a non-negligible influence on personality, such variables should be taken into account. We think that complexity should be added gradually to the model, by considering new variables – and upgrading the model itself – but only after adequate testing.

Although we found significant support for the hypotheses we formulate here, the APM is a novel theoretical model. As a consequence, it will need to be thoroughly empirically validated, especially through the operationalization and measurement of each dimension. In particular, we can consider the main hypotheses formulated by the APM – (H1) The existence of seven dimensions of attachment, each with specific characteristics; (H2) The correspondence between these dimensions and specific caregiving features; (H3) The role of the dimensions in personality and psychopathology – and suggest three different strategies (that we are already implementing) to test them.