This article shows which online message patterns signal manipulation, coercive control, or fraud — and how to shut them down safely.
What this article covers
People love the idea that one creepy text can expose a psychopath.
Reality is harder than that.
You cannot diagnose a personality disorder from a message thread. Antisocial personality disorder requires assessment by a doctor and referral to a psychiatrist or clinical psychologist, and the strongest research on psychopathy in online language shows correlations, not a magic detection test. What you can do is spot repeated patterns linked to manipulation, coercive control, fraud, and exploitation early enough to protect yourself.
Coercive control is not one behaviour or incident, but a pattern of controlling behaviour.
That is the right frame for this topic. Do not chase labels. Track patterns. Protect yourself fast.
Start with the hard truth
The internet strips away tone, facial expression, and context. That makes people overconfident. They see a few cold replies, a few lies, or a few dramatic swings and jump straight to “psychopath.”
That leap is weak.
The defensible position is simpler: a chat log may reveal risk, but it does not deliver a clinical diagnosis. If the pattern points to manipulation, pressure, fear, fraud, or control, you do not need a perfect label to act.
What the research actually found in online language
A Stanford and UBC-linked study examined archived emails, SMS messages, and Facebook messages against psychopathy scores. People with higher psychopathy scores showed more psychological distancing, wrote less comprehensible text, used more interpersonally hostile language such as anger and swear words, referred less often to their conversation partner, and were less likely to adjust their language across media types.
That matters, but keep it in proportion. The study did not prove that random readers can reliably diagnose “a psychopath” from a few messages. It showed that some psychopathy-related traits can leave linguistic traces in natural online communication. That is useful as context, not as a verdict.
So yes, language can leak something. But no, one odd phrase, one cold reply, or one angry outburst is not proof. The stronger signal is a cluster of repeated behaviors over time.
The message patterns that matter more than the label
The biggest danger online is not whether you can clinically identify psychopathy. It is whether you can recognize unsafe dynamics before they escalate.
1) Too much too soon
If someone becomes intensely romantic early, floods you with attention, and creates a “too much too soon” bond that makes boundaries harder to hold, that is a real red flag. eSafety notes that love bombing can be used to force a strong emotional bond so the person can influence and manipulate you later.
2) Hot-cold swings that keep you unstable
When someone flips unexpectedly between affection and distance to make you feel guilty, anxious, or desperate to “fix” things, eSafety says that pattern is likely controlling behavior. This matters more than a single sweet message or a single angry one.
3) Gaslighting
Gaslighting is not just lying. It is making you doubt your memory, judgment, worth, or sense of reality. In tech-based coercive control, that can include twisting facts, denying obvious behavior, or making you feel crazy for noticing a pattern that is actually there.
4) Harassment dressed up as “care”
Repeated calls, constant check-ins, demands for immediate replies, messaging your friends or family to verify your whereabouts, or expecting your passcodes are not signs of devotion. They are classic control behaviors.
5) Pressure to prove your love
If someone pushes for nudes, sexually explicit photos, passwords, private details, or financial sacrifices to “prove” trust, treat that as danger. eSafety explicitly warns that intimate images can later be used for blackmail or control.
6) Fast intimacy plus money, secrecy, or identity demands
This is where coercive control and romance fraud often overlap. Scamwatch, the FTC, and the FBI all warn about people who move fast, shift chats off-platform, invent reasons they cannot meet, push secrecy, isolate you from people you trust, and eventually ask for money, bank details, crypto transfers, accounts, or personal material that can be used against you later.
Healthy interest vs dangerous message patterns
Use this as a filter, not a diagnosis test.
| Pattern | Lower-risk behavior | Higher-risk behavior |
|---|---|---|
| Attention | Interest grows at a normal pace | Instant intensity, constant contact, “soulmate” pressure |
| Boundaries | Respects “not yet” | Pushes for passwords, nudes, money, secrecy, or immediate replies |
| Conflict | Can discuss disagreement without destabilizing you | Denies reality, flips hot-cold, guilt-trips, threatens to leave |
| Transparency | Explains delays clearly | Profile mismatches, excuses pile up, avoids meeting or video |
| Safety | Leaves you calmer and clearer | Leaves you confused, rushed, isolated, or afraid |
This table is a synthesis of the peer-reviewed language study and official guidance on coercive control and relationship scams.
How to test your suspicions without creating more risk
Do not try to “outsmart” a manipulator with drama. Use plain, low-risk checks.
- Slow the pace. Pressure thrives on speed. A safe person can handle ordinary boundaries. A risky one often escalates when you delay, refuse secrecy, or stay on-platform.
- Keep the chat where it started. Moving fast from a dating app or social platform to private messaging is a known scam red flag.
- Check consistency. Ask simple factual questions over time. Profile mismatches, contradictory details, and endless excuses matter.
- Ask for ordinary verification, not deeper intimacy. Repeated refusal to meet or video chat, especially with rotating excuses, is a warning sign.
- Search the identity. Search the name with words like “scam,” and run a reverse image search on the profile photo.
- Never send money, bank details, identity documents, or intimate images to someone you only know online. That is where “suspicious” turns into “damaging.”
How to cut them off and stay safe
If the pattern is bad enough, do not argue over the label. Act on the risk.
Preserve evidence first
People often want to delete everything immediately. eSafety’s advice is the opposite: preserve evidence first. Save the platform name, URL, usernames, profile links, dates, times, and screenshots before blocking or deleting, because you may need that evidence for the platform, police, or an online-safety complaint.
Report, then block
If this is abuse, extortion, image-based abuse, stalking, or serious harassment, report it to the platform and, where relevant, police or your national online-safety authority. If it is a romance scam, the FTC and FBI both advise stopping communication and reporting it.
Lock down your accounts
Turn on 2-step verification or 2FA on your important accounts. The UK’s National Cyber Security Centre says this makes it harder for criminals to access your accounts even if they know your password.
Check for control points
Review location sharing, connected devices, saved sessions, backup emails, password resets, and any app or account the other person may know about. Tech-based coercive control often includes tracking, monitoring, and device or account access.
Tell someone real
Talk to a trusted friend, family member, bank, platform, or relevant support service. Scamwatch and the FTC both recommend bringing in another person, because manipulators work best in secrecy and speed.
The blunt bottom line
The internet is full of lazy advice about “spotting psychopaths.” Most of it is overconfident.
The real skill is not diagnosing a stranger from one chat. The real skill is recognizing repeated patterns: distancing, hostility, control, gaslighting, pressure, secrecy, excuses, isolation, and money demands. Those patterns are enough to tell you the situation is unsafe.
Do not wait for certainty that will never arrive. If the messages leave you rushed, confused, cornered, or financially exposed, treat that as the signal. Preserve evidence. Report it. Secure your accounts. Cut contact. That is how you stay safe online.