Workplace Harassment Defined: What Power, Sexual, and Maternity Harassment Actually Mean — and Why the Lines Are Easier to Cross Than You Think

When most people think about workplace harassment, they imagine obvious, extreme behavior: a manager screaming abuse, explicit sexual advances, a supervisor openly sabotaging a returning parent. These are real, and they represent the most severe end of a spectrum.

Research on workplace harassment consistently finds most incidents are not extreme — they're the cumulative effect of behaviors that each feel marginal to the perpetrator and significant to the recipient.

Understanding what each type of harassment actually encompasses, in legal terms and in practice, is the foundation for accurate self-assessment.

→ Take the free CommStyle Assessment

Power Harassment: Definition and How It Typically Occurs

Japan's Ministry of Health, Labour and Welfare, under the Power Harassment Prevention Act, defines power harassment as conduct that simultaneously takes advantage of superiority in workplace relationships, goes beyond what's necessary and appropriate in business, and harms the working environment of workers.^[1] This definition was formalized into law under the 2020 Power Harassment Prevention Act, which for the first time made prevention measures mandatory for employers.^[2] All three conditions must be present.

Six behavioral categories are identified: physical assault, psychological attack (verbal abuse, humiliation), isolation, over-demand (unachievable targets), under-demand (tasks far below capacity as punishment), and invasion of privacy.

Research on perpetration consistently identifies the same pattern: most perpetrators didn't intend to harass — they intended to motivate performance or communicate urgency.

The mechanism that makes this gap possible: the power differential means the same intensity of feedback is experienced differently depending on which side of the relationship it comes from.

A tone that feels like appropriate firmness to the manager is experienced as threatening coercion by a subordinate with far less power to push back.

A single harsh comment may not constitute harassment.

A pattern of harsh comments, particularly directed at the same individual or in front of others, typically does — research and the legal framework distinguish between isolated incidents and systematic patterns, and systematic patterns are what most commonly generate formal complaints.

Sexual Harassment: Definition and How It Typically Occurs

Under Japan's Equal Employment Opportunity Act, employers are required to prevent sexual harassment, defined broadly as conduct of a sexual nature that either affects employment conditions based on how the employee responds, or creates a hostile work environment.^[3] Specific behaviors include questions about romantic relationships, comments about physical appearance beyond what's professionally relevant, sexual jokes in team settings, and unwanted physical contact.

Research consistently documents that most sexual harassment occurs between individuals who already know each other, in team environments where casual interaction has normalized and the boundary between collegial warmth and inappropriate conduct has shifted gradually.

This "normalization creep" pattern is particularly relevant for self-assessment — behaviors that have become habitual and haven't generated obvious objection may still be experienced as harassing by recipients who don't feel able to say so directly.

The absence of pushback isn't reliable evidence of acceptance.

Research finds only about half of employees who experience harassment report it formally, for consistent reasons: fear of retaliation, belief nothing will change, and uncertainty about whether the behavior "counts."^[5] The gap between actual and reported harassment means the absence of complaints isn't reliable evidence of the absence of harassment.

Maternity Harassment: Definition and How It Typically Occurs

Under Japan's Child Care and Family Care Leave Act, employers must prevent harassment related to pregnancy, childbirth, and parenting responsibilities.^[4] Prohibited conduct includes negative statements about pregnancy or leave decisions, pressure to forgo protected leave, and disparaging comments about the impact of parental accommodation on the team.

Research documents a specific cognitive pattern distinguishing this from other harassment types: the behavior is typically framed by the perpetrator as legitimate operational frustration rather than discrimination.

Statements like "we're all working harder to cover for their leave" feel, to the person making them, like honest descriptions of a real challenge. In the context of the employment relationship and power differential, they constitute maternity harassment.

The distinguishing question is not whether the operational burden is real, but whether the expression of that burden creates a hostile environment or generates pressure to forgo protected rights.

The Common Pattern Across All Three Types

Research on all three types identifies a single underlying pattern: the perpetrator's subjective experience of their behavior differs substantially from the recipient's experience, and the behaviors that create the most harm are rarely the behaviors that felt most significant to the perpetrator in the moment.

This pattern explains why awareness training alone doesn't reliably change behavior — the people who most need to change already believe, in most cases, they're not harassing anyone. They're not wrong that their intent isn't to harm; they're wrong that their impact isn't harming.

Behavioral self-assessment probes this gap by asking not "do you intend to harass?" but "how often do you actually engage in the specific behaviors that constitute harassment, regardless of intent?"

The Role of Power in All Three Types

All three harassment types involve a power differential as an explicit or implicit element. Power harassment is definitionally built on it. Sexual harassment is shaped by it — research consistently finds harassment is more severe and damaging when there's a career-related power differential.

Maternity harassment is enabled by it — the employee who depends on the employer's goodwill is less able to push back against operational pressure.

Understanding power differentials is central to accurate self-assessment. The same behavior feels different, and has different effects, across different power configurations — a self-assessment that probes behavior without reference to power context will systematically underestimate risk.

→ Take the free CommStyle Assessment

Frequently Asked Questions

Does harassment require intent?

No. Legal findings of harassment are based on the impact of conduct on the recipient, not the intent of the perpetrator. Research and legal frameworks consistently hold that "I didn't mean to harass" doesn't determine whether harassing conduct occurred.

This is why behavioral self-assessment, examining what you do rather than what you intend, is the appropriate framework for self-reflection.

What if a subordinate seems comfortable with behavior I'm not sure about?

Research on power differentials consistently documents that subordinates frequently don't express discomfort with harassing behavior to the perpetrator, because doing so feels unsafe or unlikely to change the situation.

The absence of expressed objection isn't reliable evidence that conduct is welcome or acceptable.

What counts as "going beyond what is necessary" in power harassment?

This is a judgment-based standard depending on context. Key questions: would a reasonable person in the field find this level of correction proportionate to the issue? Is this consistent with how you treat other employees in similar situations?

Is the manner and setting appropriate — private rather than public, focused on the work rather than the person? When the answers are uncertain, the behavior is likely in or near the risk zone.