Old and interesting #2: The Great Cabbage Hoax
The anatomy of a fake news story, 1950s edition
[This post is a selective summary of and a reflection on the paper The great cabbage hoax: A case study. The paper was published in 1965 in the Journal of Personality and Social Psychology, a leading social psychology journal.]
On the 16th of November 1964, The National Observer, a weekly newspaper, reported the following: “The Ten Commandments contain 297 words. The Bill of Rights is stated in 463 words. Lincoln's Gettysburg Address contains 266 words. A recent Federal directive to regulate the price of cabbage contains 26,911 words.”
This excerpt about regulating the price of cabbage seems to confirm what many people had suspected for a long time: the state is becoming overly bureaucratic and dysfunctional. I can imagine many of you (myself included) nodding in agreement and starting to think about ways to finally curb that ever-expanding state power. But before we get too immersed in this rather satisfying mental exercise, let’s deal with a small inconvenience: the news was a false rumor. There was no federal directive regulating the price of cabbage around that time.
The last time the United States implemented price controls before 1964 was during the early 1950s, specifically during the Korean War (1950-53). During that period, it was the responsibility of the Office of Price Stabilization to set maximum prices on certain products, but cabbage was not one of them.
It’s unclear when this false rumor started, but it appeared in widely circulated written materials as early as 1951. Despite repeated efforts to debunk it, the cabbage hoax persisted for over a decade. Not only that: it also gave rise to other false rumors of a similar flavor.
The editor of a weekly newspaper came across the cabbage story in late 1951 and decided to reprint it. Unlike other editors who reprinted it without checking it for accuracy, thereby inadvertently spreading the false rumor, he checked it for accuracy. He learned that the cabbage price directive didn’t exist. But the story was too good to pass up, so he searched for another product in the Machinery Regulation, one of the most comprehensive directives that the Office of Price Stabilization had ever issued. Eventually, he stumbled upon what he needed: manually operated foghorns.
The foghorn became the new cabbage. Like in the case of the cabbage hoax, the relevant text was featured alongside short yet foundational texts like the Ten Commandments. It read as follows: “12,962—Words in OPS [Office of Price Stabilization] order establishing the ceiling price of manually operated foghorns and other manufactured items.”
The order included practically every kind of machinery (close to 380 in total). I couldn’t find the order, but I suspect the manually operated foghorn wasn’t the most central item within it. In other words, the editor could have selected many other manufactured items that would have conveyed the scope of the regulation more accurately. But all the alternatives probably lacked one crucial aspect that made the foghorn a great choice: entertainment value. The more niche the item, the starker the contrast and the funnier the story. When you are in the business of selling your publication, this type of story certainly doesn’t hurt.
The story doesn’t stop here. The foghorn hoax took on a life of its own. Other newspapers reprinted it with seemingly innocuous modifications that proved quite consequential. The ‘foghorns and other manufactured items’ part first became ‘foghorns and related manufactured items.’ The careful reader will notice that the second version covers much fewer items. While the first version could potentially include all manufactured items, the second version only covers a niche item and its similarly niche related items. The next version went a step further: the price directive was allegedly about ‘manually operated foghorns, etc.’ Finally, even the ‘etc.’ was dropped from the text. It didn’t take many iterations to complete the journey from a factually accurate yet misleading claim to an outright false claim.
The way the foghorn story underwent gradual changes reminds me of dialect continuums. Neighboring dialects may be mutually intelligible, but as you move farther away from a particular area, the dialects may become increasingly different to the point of being unintelligible. Similarly, the text of the foghorn story became distorted every step of the way. While we see the similarity between a version and the one preceding it, the difference between the original and the final versions seems abysmal.
A joke from Eastern Europe’s communist era illustrates this point well: "Is it true that they're giving away Mercedes in Moscow?" "Yes, it's true," reports the Yerevan Radio, "with one small correction: it's not Moscow, but Leningrad; not Mercedes, but Zhiguli; and they're not giving them away, but taking them."
Which rumors spread?
This rumor spread quickly because it was about something important, namely the bureaucratic and dysfunctional state. Additionally, the information surrounding the situation was ambiguous or uncertain, as specific pricing directives tend to fly under most people’s radars. As post-hoc and general explanations typically do, this explanation seems to fit the facts of this case perfectly. The real question, however, is not whether we can come up with a plausible-sounding narrative to explain why specific rumors took off. We can always do that. It is whether our theory makes actionable predictions that we can test on future data.
Manufacturing rumors that consistently succeed is extremely difficult—if not impossible. For instance, secret services, political activists, and estranged lovers have been in this business for a long time, and their track record is mixed at best. Similarly, if we look at the rumor’s modern cousin, the viral tweet, the conclusion remains unchanged.
As in many other domains, there is an asymmetry: The absence of relevant factors leads to failure, but the presence of relevant factors doesn’t guarantee success. For example, a rumor about the party affiliation of a president will almost certainly not take off because the information can be easily checked. In contrast, if it’s about the same president’s secret lovers, tooth brushing habits, or mental health treatment, the rumor has a larger chance of succeeding because reliable and verifiable information on these topics is difficult to come by. In absolute terms, however, the chance of success is still very small even when the relevant factors are present. If we consider one thousand a priori feasible rumors, we can be relatively confident that a handful of them will become successful, yet we cannot predict which ones.