How do I know if someone new is lying to me?

Usually you can’t tell from how someone acts in the moment, and neither can anyone else, so stop auditing their face and start watching whether their story holds up over time. If you were fooled before, that isn’t proof your judgment is broken. It’s proof you were using the tools everyone uses, and those tools are weaker than we’re told.

Why can’t I just read it on their face?

Picture it. You’re 34, three dates in with Theo. He says he’s been single “about a year.” Then, over dinner, he mentions a trip to Lisbon last spring “with a friend,” and his eyes drop to the menu. Your stomach does the thing. The last person you dated said “a friend” a lot, too.

Here’s the problem with the menu glance. When researchers asked people in 58 countries how they could tell someone was lying, the most common answer everywhere was the same: liars look away.1 It’s the planet’s favorite lie-detection rule. It just isn’t a reliable one.

When psychologists Maria Hartwig and Charles Bond pooled decades of lie-detection studies, they found that people’s poor accuracy isn’t mainly about watching the wrong cues. It’s that the behavioral differences between liars and truth-tellers are weak to begin with.2 There’s no secret tell to learn, because mostly the tell isn’t there.

So if someone you trusted lied to you and you missed it, you didn’t fail a test everyone else passes. You were working with signals that are faint for everybody.

Why does it feel harder to tell with someone I like?

Because liking someone changes how you look at them. Deception researchers Steven McCornack and Malcolm Parks proposed that as people get closer, they grow more confident they can read each other, that confidence tilts them toward believing what they hear (a “truth bias”), and accuracy drops.3

That isn’t a flaw. Believing the people you’re getting close to is how closeness happens at all. But it explains a familiar feeling after a lie comes out: “I was so sure.” Certainty isn’t the same thing as accuracy, in either direction. The alarm you feel at Theo’s menu glance and the warm certainty you once felt with someone else can both be wrong.

So how do people actually find out?

Mostly later, and mostly not from body language. When Timothy Levine and colleagues asked people to recall a lie they had actually caught, most of the stories involved evidence or a confession, and the discovery often came weeks, months or even years after the lie.4

That might be the most useful research finding here, because it changes the job. You’re not trying to catch a lie in real time across a dinner table. You’re letting time do what it does: stories either keep their shape or they don’t.

Some things that tend to surface with time:

• Details that stay put. The Lisbon friend has a name by date six, and the name is still the same on date ten.
• Words that match actions. “I’m not seeing anyone else” next to plans that only ever happen on weeknights. One odd thing means little. A pattern is worth a question.
• A life with other people in it. Over a few months, someone being straight with you usually lets you meet some of the people in their stories.
• How they take a plain question. Not only the answer. Whether they can answer without getting irritated that you asked.

Interview research points the same way. Across studies of the so-called verifiability approach, people telling the truth tended to include more details that could be checked, like a named person or a specific place, than people who were lying.5 You’re not interviewing anyone. But a story with real names and places in it, which stays the same over time, is worth more than any glance.

None of these tells you what someone intends. They’re facts you can notice without turning into a detective.

Is it different when you met online?

A little, because you start with less shared context: no mutual friends, no office, no overlapping history to check a story against. Most dishonesty you’ll meet early on is small, like a softened age or an “about a year” that was really four months. Still, it’s a good reason to let the people in someone’s life become visible before you get in deep, so their stories have somewhere to land.

How do I trust my gut again after being fooled?

Start by pulling apart four things the next time your stomach drops:

• What happened: “He said he went to Lisbon with a friend and looked at the menu.”
• The reading: “He’s hiding who the friend was.”
• The fear: “I’m being fooled again, and I’ll miss it again.”
• The next step: ask about Lisbon, warmly, and let the answer sit for a few weeks.

Often the first line is thin and nearly all the weight is in the third. That’s not a reason to ignore your gut. It’s a reason to give it a job it can actually do: noticing when things don’t add up over time, rather than reading a single glance.

It also helps to change what “trusting yourself” means. It can’t mean “I’ll never be fooled,” because nobody gets that guarantee. It can mean “if something doesn’t line up, I’ll notice, I’ll ask, and I’ll act on what I learn.” That’s a promise you can keep.

If the distrust has become constant, if you’re checking someone’s profiles at 1 a.m. or it’s bleeding into your sleep and work, that’s worth bringing to a licensed mental-health professional. Not because something is wrong with you, but because it’s a lot to carry alone.

What to do this week

• Retire the face test. When you catch yourself reading eye contact, name it (“there’s the menu glance again”) and move on.
• Ask one easy, curious question about the detail that bugged you. Then let it rest, and notice how they answered, not just what.
• Keep a light “does it hold” note. Not a file on them. Two or three things they’ve told you, so you’ll notice if they shift.
• Let people in. Suggest something with one of your friends, and say yes when they offer the same.
• Decide your line while you’re calm. What would you need to see to keep going, and what would make you step back? Writing it down now is a favor to you on date eight.

What doesn’t the research show?

• Lie-detection studies mostly use strangers in labs, often judging short clips. Real dating gives you more context and more time.
• The truth-bias idea is a model with research behind it, not a measurement of you or the person you’re seeing.
• The caught-lie findings rely on people’s memories of lies they discovered, which can be selective.
• The checkable-details research comes from structured interviews, often in legal or insurance-style settings, not from dates.
• Nothing here can tell you whether a particular person is lying. It can tell you where to look: at what holds up over time.