The Pygmalion Effect Through the Lens of CDT

I was reading an article about the Pygmalion Effect and found myself asking an awkward question. If, through the lens of Constructed Development Theory, we cannot construct another person directly or completely, how can our expectations supposedly construct their performance?

The article was Pygmalion Effect: What is it and how to use it as a manager?, published by Management 3.0. Its central message is familiar: expect more from people and they are more likely to perform well; expect less and they may fulfil that lower expectation. For managers, mentors and teachers, the lesson appears both optimistic and practical. Our expectations influence other people, so we should become aware of those expectations and use them positively.

There is something important in that claim. There is also something psychologically incomplete.

The problem is not necessarily whether interpersonal expectations have consequences. They clearly can. The problem is the way the mechanism is described. Conventional accounts often move too easily from I expect something of you, to I behave differently towards you, to you become what I expected. In that movement, influence quietly becomes determination, prediction becomes causation and the person receiving the expectation becomes strangely passive.

CDT does not invalidate the Pygmalion Effect. It exposes the incomplete psychology beneath its usual explanation.

What the Management 3.0 article gets right

The strongest part of the article is not its simple five-step self-fulfilling prophecy cycle. It is the four-factor account of how expectations may be communicated through behaviour. Drawing on work associated with Rosenthal and colleagues, it suggests that teachers—and, by extension, managers, mentors and clinicians—may treat people from whom they expect superior performance differently in four areas:

1. Climate: they create a warmer social and emotional environment, often through subtle verbal and non-verbal behaviour.
2. Feedback: they provide more frequent, differentiated or useful information about performance.
3. Input: they offer more material, greater complexity or more demanding developmental opportunities.
4. Output: they provide more time, space and opportunity for the person to respond, experiment and demonstrate what they can do.

These four factors matter because they provide a behavioural mechanism. The expectation does not float telepathically from one mind into another. It changes the conduct of the person holding it. That conduct changes the interpersonal environment. The changed environment then becomes part of the information available to the other person.

We do not construct the person

The opening example in the Management 3.0 article describes a mentor who receives advance information about two mentees. The first is presented as brilliant and successful. The second is described as someone who struggles and produces poor-quality work. Before meeting them, the mentor has already predicted which one will learn faster.

Through a CDT lens, the mentor has not constructed either mentee. The mentor has constructed an internal model of each mentee from a supplied story.

That distinction is foundational. I do not have unmediated access to you. I receive information about you: what you say, what you do, how others describe you, what I remember, what I notice and what I fail to notice. That information passes through my existing Thinking Style, Cognitive Intentions, previous experience, cultural assumptions, emotional state and current context. What emerges is not you. It is my construction of you.

This construction may be relatively well calibrated. It may be badly mistaken. It may contain useful observations alongside projections and inherited assumptions. What it cannot be is exhaustive or identical to the construction you have of yourself.

It would therefore be too strong to say that we can never construct another person “accurately” in any sense. We can make better and worse predictions about observable behaviour. We can update our understanding through sustained contact. Some interpersonal judgements are clearly more informed than others. Accuracy is a matter of degree, evidence, context and purpose.

What we cannot do is construct another person directly, completely or independently of our own constructive process.

There are two constructing systems, not one sculptor

A CDT account of the Pygmalion Effect must contain two active construction processes.

The manager’s construction

1. The manager receives information about the employee.
2. Cognitive Intentions select, filter, generalise and organise that information.
3. The manager constructs an expectation about ability, potential or likely behaviour.
4. The manager may or may not become aware that this expectation is a construction.
5. The expectation influences available Choice and subsequent Response.
6. The manager creates a particular climate and distributes feedback, challenge, attention and opportunity accordingly.

The employee’s construction

1. The employee encounters the manager’s changed behaviour as new input.
2. The employee filters that behaviour through their own Cognitive Intentions, history, expectations and context.
3. The employee constructs meaning: support, trust, pressure, manipulation, indifference, opportunity, threat or something else.
4. The employee may or may not become aware of that construction.
5. The employee exercises whatever Choice is available within the situation.
6. The employee responds, and that Response may affect performance.

The employee’s Response then becomes further input for the manager. The manager may revise the original expectation—or selectively interpret the Response as proof that the original judgement was right. The process is recursive.

In its shortest form:

This is not one person mentally creating another. It is an interaction between two constructing systems, occurring within a relationship shaped by unequal roles, resources, histories and power.

The recipient is not passive material

The Pygmalion metaphor comes from the sculptor who brings his idealised creation to life. It is rhetorically powerful, but psychologically misleading. It implies an active creator and passive material. The manager imagines; the employee becomes. The teacher expects; the pupil fulfils. The observer creates reality.

CDT cannot accept that one-directional account.

The employee may accommodate the manager’s expectation, but they may also:

• resist the expectation;
• react against it;
• reinterpret it;
• fail to notice it;
• experience it as pressure rather than encouragement;
• recognise the label and deliberately challenge it;
• perform differently because concrete opportunities were added or removed;
• or remain largely unaffected.

None of this means expectations are irrelevant. It means they alter probabilities and conditions rather than dictating identity or Response. The recipient remains an agent, although their available Choices may be widened or constrained by the relational environment.

This point is especially important in organisations. Managers control access to information, developmental assignments, visibility, resources, feedback, tolerance for mistakes and opportunities to speak. An employee may appear to fulfil an expectation simply because the organisation materially equipped one person to succeed while withholding the same conditions from another.

That is not magic. It is not the expectation reaching into the employee and changing their essence. It is the distribution of affordances.

Where the familiar account becomes deterministic

The Management 3.0 article presents a five-stage cycle:

1. we develop an expectation;
2. we express it verbally or non-verbally;
3. others adjust their behaviour to match ours;
4. the expectation becomes reality;
5. the apparent confirmation strengthens the original belief.

The first two stages are plausible. The fifth captures a genuine risk of reinforcement. The conceptual problem lies in stages three and four. Other people do not simply adjust themselves to match us, and our expectations do not reliably become reality—let alone shape a situation into the “exact” form predicted.

Something more conditional is occurring:

Our construction may alter our behaviour. Our behaviour may change some of the other person’s inputs, opportunities or constraints. The other person constructs those conditions and responds from within their own developmental capacity and context. Their Response may move in the expected direction, in the opposite direction or nowhere meaningful at all.

When the expected outcome does occur, we still have to establish why. Was the original expectation accurate? Did we create better material conditions? Did the employee respond to our altered behaviour? Did we notice only confirming evidence? Did some unrelated factor produce the outcome?

Without answering those questions, “my expectation came true” tells us very little.

Four distinctions the popular account collapses

Any serious account of the Pygmalion Effect must separate four processes that are frequently treated as though they were the same thing.

Only the fourth process is properly a self-fulfilling Pygmalion Effect. The others may coexist with it, but they are not interchangeable.

This distinction also explains why an inaccurate initial expectation is so important to the strongest causal claim. If a manager identifies an existing difference and later performance reflects that difference, the expectation may have been accurate rather than self-fulfilling. Classic experiments attempted to isolate the causal effect by randomly assigning supposedly high-potential labels. Because the labels were not based on actual measured differences, any systematic change could more plausibly be attributed to the expectations and the behaviours they produced.

However, demonstrating that an induced expectation changes outcomes does not show that people are infinitely malleable, nor that ordinary managerial expectations have the same power. It shows that, under some conditions, a false construction can change behaviour sufficiently to alter another person’s environment.

Can managers really “control” their expectations?

The article advises leaders to “be in control” of their expectations and to review what they expect of their colleagues. The developmental intention is sensible. The psychology is too neat.

We may not consciously choose the first construction that appears. Habituated Cognitive Intentions operate rapidly. Existing categories, emotional associations, organisational narratives and previous experience can organise perception before reflective Awareness enters the process. A manager may discover that they have already classified someone as talented, difficult, safe, unreliable, leadership material or “not quite ready.”

CDT does not require the impossible demand that no automatic construction occur. It asks whether the person can become aware of the construction as a construction.

The developmental movement is therefore not:

Choose the correct expectation and make it come true.

It is:

Notice that the expectation is your construction, examine what produced it, and choose how much authority to give it in your Response.

A manager may become aware of a pattern but still lack the contextual power, emotional regulation or developmental capacity to change it immediately. Awareness does not guarantee Choice. That distinction is central to CDT. Nevertheless, without Awareness, the habituated construction is much more likely to pass directly into behaviour.

No label is neutral

The article recommends avoiding comparative and superlative descriptions such as “brighter” or “the best.” That is useful. It then proposes finding alternative, supposedly neutral labels for team members, offering “Padawan” as an example.

But there is no cognitively neutral label.

Every label selects some features and ignores others. Every category carries associations. “Padawan” constructs a learner, but it also constructs a hierarchy, a master, a developmental deficit and a particular relationship to authority. Replacing one label with another does not remove the act of construction.

The deeper CDT move is not to search endlessly for the perfect neutral category. It is to become aware that categorisation is happening, to hold the resulting construction provisionally and to remain capable of updating it.

Language matters because it can stabilise a temporary observation into an identity. “This report contains errors” is not the same construction as “she is careless.” “He has not yet demonstrated this capability” is not the same as “he lacks potential.” Behavioural descriptions remain open to context and revision. Identity labels invite us to find a person wherever the label tells us to look.

Positive expectations are not automatically harmless

The popular leadership lesson is usually to expect the best from everyone. It sounds humane, and it may encourage managers to distribute attention and opportunity more generously. But positive expectations are not intrinsically accurate, ethical or beneficial.

Unrealistically high expectations can:

• create performance anxiety and fear of disappointing the leader;
• encourage overwork or concealment of difficulty;
• produce opportunities for which the person has not been adequately supported;
• cause a manager to ignore evidence that the employee wants a different developmental path;
• allocate scarce resources according to optimism rather than need or evidence;
• turn confidence into pressure, obligation or control;
• and make the employee responsible for preserving the manager’s preferred story about them.

The ethical goal is not uniformly high expectation. It is equitable access to developmental conditions, evidence-sensitive judgement, openness to revision and respect for the other person’s agency.

Low expectations can clearly restrict opportunity. High expectations can expand it. But either can become a colonising construction when the manager stops encountering the person and begins managing the story they have created about the person.

What the evidence allows us to say

The original Pygmalion in the Classroom research became enormously influential, but its methods, statistical treatment and interpretation have also been heavily contested. The broad research literature does not support the simple claim that expectations routinely manufacture large changes in intelligence or performance.

Jussim and Harber’s review of 35 years of research concluded that classroom self-fulfilling prophecies do occur, but that they are usually small, do not generally accumulate dramatically over time and may often dissipate. They also argued that teacher expectations predict outcomes substantially because the expectations are accurate—not merely because they are self-fulfilling. More powerful effects may occur selectively, including among members of stigmatised groups.

Raudenbush’s meta-analysis of 18 experiments found that induced teacher-expectancy effects became smaller when teachers already knew their pupils. False information had more power when introduced before or very early in the relationship.

This is particularly interesting through a CDT lens. Once a teacher has constructed a relatively stable model of a pupil through repeated interaction, an externally supplied label may be filtered against that established construction and rejected. Where there is little prior contact, the label can become an early organising frame through which subsequent information is selected and interpreted.

Later intervention research also suggests that expectancy-related practice can be changed. A systematic review and meta-analysis of teacher-expectation interventions found small-to-moderate improvements in teacher expectations and student achievement. However, the authors noted the limited number of studies that measured changes in both expectations and achievement, making the complete causal pathway difficult to establish.

None of this renders the Pygmalion Effect fictional. It renders it conditional.

Expectations can matter, particularly when they change concrete behaviour and opportunity. They are not spells. They do not override the recipient’s constructive process. They are neither always false nor always causal. Their consequences depend on context, prior knowledge, power, behaviour, interpretation and the agency of everyone involved.

A CDT reconstruction of the Pygmalion Effect

The usual definition says that one person’s expectation of another can become a self-fulfilling prophecy. A CDT-compatible definition needs to be more precise:

This definition deliberately refuses several common exaggerations.

• It says probabilistic, because the outcome is not determined.
• It says recursively mediated, because both people continue constructing and responding.
• It locates the expectation within the observer’s construction, rather than treating it as knowledge of the other person.
• It identifies changed inputs, opportunities and constraints as the mechanism.
• It preserves the recipient’s own construction and Choice.
• It focuses on performance in a context rather than claiming that one person creates another person’s identity or potential.

Through this reconstruction, the fact that we cannot construct another person directly is not a fatal problem for the Pygmalion Effect. Inaccurate interpersonal construction may be the very event that initiates the process. The expectation can have consequences because we act from it, not because it accurately contains the person.

The stronger challenge is epistemic and ethical: are we aware that we are acting from a construction? What evidence produced it? Which Cognitive Intentions organised it? What possibilities are we making available or unavailable because of it? How are power and organisational structure amplifying it? What evidence would persuade us to revise it?

What this means for leaders

A leader informed by CDT should not try to become a benevolent Pygmalion who sculpts other people into improved versions of themselves. That still centres the leader as creator and reduces colleagues to developmental material.

A more deliberate practice would involve the following:

1. Separate the person from your model of the person. Use language such as “my current construction is…” to preserve epistemic humility.
2. Trace the construction. Ask what information, hearsay, previous interaction, category, emotion or Cognitive Intention produced the expectation.
3. Describe behaviour before assigning identity. Keep observations contextual, specific and open to revision.
4. Audit the distribution of conditions. Compare who receives warmth, challenge, feedback, patience, visibility, information and second chances.
5. Do not confuse unequal opportunity with unequal potential. Performance may reflect what the system made possible.
6. Invite the other construction. Ask colleagues how they understand the situation, what they want and what your behaviour communicates to them.
7. Look actively for disconfirming evidence. Decide in advance what would cause you to update the expectation.
8. Hold positive expectations without imposing an identity. Offer opportunity and support without requiring the person to inhabit your preferred story.
9. Distinguish Awareness from Choice. Noticing a bias is developmentally important, but further work may be required before a different Response becomes reliably available.
10. Review the recursive loop. Ask whether the behaviour now being judged is partly a Response to conditions you helped create.

This is a more demanding leadership position than simply “expect the best.” It requires the leader to examine not only the content of an expectation but the construction process beneath it and the organisational consequences produced by acting from it.

Closing thought: from sculpting people to constructing conditions

The Pygmalion Effect survives a CDT analysis, but the mythology surrounding it does not.

We do not think another person into becoming. We do not accurately contain them in our expectations. We do not control the meaning they make from our conduct. We do not guarantee that positive expectations will produce positive results.

What we can do is construct an account of someone, act from that account and thereby change the conditions they encounter. Those conditions may influence how they construct their own possibilities and Responses. Their Response then returns to us as fresh information, which we may use to revise our construction—or misuse to confirm what we had already decided.

The developmental responsibility lies in recognising that loop.

The leader is not the sculptor of another person. The leader is one participant—often a powerful one—in the construction of the conditions within which both people respond.

That is less magical than the traditional Pygmalion story. It is also more psychologically coherent, more ethically responsible and far more useful.