M: I need you to quickly look up a particular story that I've only heard, like, not quite anecdotally, but referenced without actually giving me details in terms of where and when it happened. Evidently, I think it was in Australia. There was someone who accused someone else of raping them. And it turned out that the reason they accused that particular person wasn't because they were anywhere near them, but because they, that person had been on the TV when the attack occurred.
G: I wasn't able to find a case matching those specific facts, and I suspect the story may have been compressed or slightly altered in retelling.
There are, however, two closely related phenomena that may have become conflated:
The victim saw the suspect's face in the media before identifying him. This is a well-documented problem in eyewitness psychology. Once someone has seen a face repeatedly—on television, in newspapers, or online—they may later experience source-monitoring error, sincerely believing they recognize the person from the crime itself rather than from the later media exposure. This issue has been discussed in numerous wrongful-conviction cases and by memory researchers such as entity["people","Elizabeth Loftus","memory researcher"]. citeturn1academia6turn0search14
The face was present during the crime because it was on television. This is rarer, but entirely plausible psychologically. If a television happened to be on during an assault, the victim could later remember a face that was genuinely present in the room—but only on the TV screen. Under trauma, the brain may later misattribute the source of that memory, leading to an honest but mistaken identification. Memory researchers have described this kind of source confusion, although I couldn't find a famous Australian case that exactly fits your description. citeturn1academia6turn0search14
One Australian case that touches on the issue is the murder of entity["people","Peter Falconio","Australian murder case"]. The defense argued that eyewitness entity["people","Joanne Lees","Falconio witness"] had seen a photograph of the suspect on the internet before identifying him, potentially contaminating her memory. The courts ultimately rejected that argument, but it illustrates the concern about post-event media exposure influencing identification. citeturn1search1turn1search4
So my guess is that the story you heard is either:
If you remember any additional detail—even something as small as:
I can probably narrow it down. The detail that the victim picked the person because they were literally on the television during the assault is distinctive enough that another clue may be all that's needed.
M: I'm fairly certain it was a rape. I can't say for certain whether it was while the assault was occurring or immediately before the assault occurred.
G: [Non-text content: thoughts; content_type=thoughts]
G: I’ll treat “Australia + rape + TV exposure/source confusion” as the core pattern and search around both legal-case wording and psychology/eyewitness-memory references, since this may be cited under a researcher’s example rather than a news headline.
G: [Non-text content: thoughts; content_type=thoughts]
G: [Non-text content: reasoning_recap; content_type=reasoning_recap]
G: You found it: Donald Thomson, Australian psychologist/memory expert.
The case is usually cited as a classic example of source confusion / unconscious transference. Thomson had been on live Australian TV discussing eyewitness-memory unreliability. A rape victim later identified him, but the rape occurred while he was on TV; she had apparently been watching the broadcast around the time of the attack and misattributed his face to the assailant. citeturn491383search1turn491383search12
A few sources say the victim was watching him when the rapist broke in; others phrase it as before the crime. So your uncertainty there is reflected in the retellings. citeturn491383search5turn491383search17
One caveat: a recent Medium post questions how well-documented the original 1975 account is, so it’s widely repeated in psychology teaching, but not as cleanly sourced as a normal court/news case. citeturn491383search6
[Attachment reference unresolved: file-family]
[Attachment reference unresolved: file-victo]
M: Is there any documentation on the 1975 account? Like an actual case file or something like that?
G: I went looking specifically for primary documentation—court records, police reports, newspaper coverage from the time, or Thomson's own contemporary account.
The short answer is: I've found no publicly available primary record.
What I did find is a bit surprising:
That absence has become noteworthy enough that some researchers have questioned the story's provenance. The skepticism is not so much that "it definitely didn't happen," but rather that an anecdote this famous has a surprisingly thin documentary trail. citeturn0search4
There are a few possibilities:
Given the number of independent textbooks that repeat it, I'm inclined to think there is some historical basis for the event. But as a historian, I would distinguish between:
The first is true. The second, at least based on what's publicly accessible today, is not.
If you're interested, I can take a deeper dive—not just a web search, but into Australian newspaper archives, early psychology journals, and Donald Thomson's own publications from the mid-1970s—to see if we can trace the anecdote to its earliest verifiable appearance. That's the sort of thing that often uncovers whether a famous story is rooted in an actual event or has become a textbook legend through repeated retelling.
M: Don't make a deep dive. It's not important enough for that at the moment. Is the actual phenomenon well documented, though, in a way that doesn't require that kind of deep dive? Like, have they come up with actual instances which might not be quite as dramatic but are just as, just as potent of a demonstration of how memory can be fucked up or overwritten by circumstances?
G: Yes. In fact, I'd go further than that: the phenomenon is among the best-established findings in cognitive psychology. The Donald Thomson story is memorable because it's so dramatic, but the underlying effect does not depend on that anecdote. It's been reproduced hundreds of times in laboratories and has shown up repeatedly in real criminal cases. citeturn0search1turn0search13
There are really three related phenomena that have overwhelming evidence behind them:
The misinformation effect. If you witness an event and are later exposed to misleading information—even something as simple as a suggestive question—that new information can become part of your memory. Elizabeth Loftus's classic work demonstrated this repeatedly. People don't simply "repeat" the misinformation; many come to sincerely remember it as part of the original event. citeturn0search13turn0search12turn0search1
Source-monitoring errors. Your brain often remembers the information while forgetting where it came from. You may accurately remember a face, a sentence, or an object but incorrectly attribute its source: "I saw him during the crime" instead of "I saw him on television," or "The officer told me that" instead of "I observed that." This is considered a normal feature of memory rather than an exotic failure. citeturn0search2turn0search23turn0search21
Unconscious transference. This is especially relevant to eyewitness identification. Someone who was merely present—an innocent bystander, another customer, someone previously seen in a mugshot—can later be identified as the perpetrator because the witness correctly recognizes the face but misremembers why it feels familiar. Laboratory studies have repeatedly demonstrated this effect. citeturn0search22turn0search21
One real-world example that's exceptionally well documented is the case of entity["people","Ronald Cotton","wrongful conviction case"]. The victim, entity["people","Jennifer Thompson-Cannino","eyewitness in Ronald Cotton case"], was absolutely convinced of her identification. She was conscientious, paid close attention during the assault, and remained highly confident. DNA evidence later proved she had identified the wrong man. Years afterward, she described that she genuinely "saw" Cotton's face in her memory of the attack even after learning he was innocent. That case became one of the strongest demonstrations that confidence and sincerity are not reliable indicators of accuracy. citeturn0search22turn0search5
The interesting shift over the last couple of decades is that researchers have become more precise in how they describe this. Earlier work sometimes left the impression that memory is simply "unreliable." More recent work argues that's not quite right. Instead, memory is highly susceptible to contamination. If you capture a witness's memory immediately, before discussion, media exposure, or suggestive questioning, it's often much more accurate than people assume. The problem is that memory is reconstructed every time it's retrieved, so every later interview, news report, or conversation is an opportunity for contamination. citeturn0search1turn0search5
One aspect I find particularly interesting—and I think it aligns with why this caught your attention—is that these errors are not lying. The person is usually reporting exactly what they experience. Their internal experience of certainty is genuine. The error isn't in honesty; it's in the brain's bookkeeping. It has stored the content but attached the wrong provenance. That distinction is one of the reasons eyewitness testimony has become much more carefully scrutinized in modern forensic psychology. citeturn0search2turn0search23turn0search1
M: Confabulation.
G: Yes—confabulation is adjacent to this, but it's not quite the same thing.
Confabulation is when the brain fills in missing information with something that feels true to the person. They're not intentionally lying; the brain is essentially improvising to create a coherent narrative. It's most commonly associated with neurological conditions (such as damage to the frontal lobes or entity["disease","Korsakoff syndrome","alcohol-related amnestic disorder"]), but milder forms can occur in healthy people when memory is incomplete.
What we've been discussing is usually classified more specifically as:
Those don't necessarily require the brain to invent anything. The face, the voice, or the detail is real—it has simply been attached to the wrong event.
Confabulation, by contrast, is more like:
"I don't actually remember what happened here, so my brain constructs a plausible sequence without realizing it's doing so."
The reason these concepts are often discussed together is that they all illustrate the same broader principle:
Memory is reconstructive, not reproductive.
People often imagine memory as a video recording that can become fuzzy. Modern psychology instead treats it more like a document that's reopened and re-saved every time you access it. Most of the time, that process preserves the document well. But occasionally it introduces edits—sometimes from other memories, sometimes from outside information, and sometimes from the brain's own attempt to fill in gaps.
So if you were looking for a broader umbrella term in conversation, confabulation isn't a bad one. If you wanted to be technically precise in the context of eyewitness identification, though, I'd reach first for source-monitoring error or unconscious transference, because those describe the specific mechanism at work rather than the more general phenomenon of memory reconstruction.