Cannabis Demographics as Reported in the Press, 1910-1919

One of the most influential theories of cannabis prohibition in the United States is the so-called “Mexican hypothesis,” which suggests that cannabis was prohibited in various states as a pretext to harass Mexican immigrants, and that anti-Mexican attitudes helped spawn the stereotypes that cannabis use led to madness, violence, and crime. Recently that theory has come under attack from various scholars, myself included. We have collectively argued that the evidence for this hypothesis is scant, that the most lurid stereotypes of marijuana’s effects actually originated in Mexico, that marijuana use was relatively rare in Mexican immigrant communities, that many of the earliest cannabis prohibitions occurred in states where there was essentially no Mexican presence and almost no discourse about Mexican “marihuana,” and that far more influential in the development of national narratives about this drug was the city of New Orleans, where the discourse was not dominated by anti-Mexican discourse or stereotypes.[1]

Nonetheless, the notion that drug prohibitions in general were a product of racism of one kind or another has been highly influential and has been used as an argument for drug policy reform since the 1970s. Clearly drug laws have been enforced in a biased and racist way, but this does not necessarily mean that this was their purpose or inspiration. We know, for example, that Black Americans are also more likely to be ticketed for traffic violations than White Americans. This of course does not mean that speeding laws were developed in order to provide a pretext for harassing the Black community. And even if we assert that race was a key factor in the origins of drug laws, what specifically do we mean by that? Was it, as some authors have argued, that drug laws were linked to minority communities in order to provide a tool for better controlling labor? Was it, as in the most commonly forwarded version of the “Mexican hypothesis,” that these laws were developed to provide the police a pretext for harassing immigrant or other marginalized groups? Was it, as others have argued, that the laws were intended to prevent the “infection” of White people by these supposedly alien practices that, had they remained confined to the marginalized, would not have been considered especially alarming? And, if marijuana was associated with madness and violence simply because, as I have argued elsewhere, those ideas came to the United States from Mexico basically fully formed, wasn’t that enough on its own to justify prohibition even if some of those calling for that prohibition were also racist? [2]

Methodology

Before we look at the data, a few comments on methodology are in order.

As with our analysis of cannabis effects, we had to make some subjective decisions in our categorization of the various demographic groups described in the early-twentieth-century press. First, for example, the stories on hashish often involved a variety of stereotypical “Oriental” characters in the rather racist, orientalist fashion of the period (if you aren’t familiar with the concept of “orientalism,” this is a good place to start). These might be described by the sources as Egyptians, or Persians, or Turks, or Indians, or “Mohammedans,” or something else (Egyptians were the most common). Because these stories were blatantly orientalist, and because the distinctions between these groups were surely relatively irrelevant to early-twentieth-century readers imbibing exotic stories of the “Orient,” I decided to consolidate all such references into the simple descriptive category of “Orientals” in quotation marks. We are of course well aware of the racist connotations of the term “oriental.” But I decided that we should utilize this term for the visualizations given that there is a huge literature on “orientalism,” and the term best sums up a racist but ubiquitous set of ideas that were both embedded in these stories and quite familiar to early twentieth century readers. Furthermore, “oriental” was not used the same way as “the N-Word” or similar slurs for other races. Someone didn’t hurl the word “oriental” at a person to insult them. It was a descriptor used in academic and popular discourse often without any conscious malicious intent. Indeed, one could argue that the word “Mexican,” uttered with a certain menacing inflection, was far more likely to be used as a slur during this period in certain parts of the country than the word “oriental.” I just don’t think there’s any other word that better sums up this collection of ideas about “non-Western” people, and hence my decision to use it here.

Second, we were often confronted by stories, particularly those that described specific “incidents” (more on this below), that were rather vague regarding the demographics of those involved. For example, some stories would feature someone with a Hispanic surname, but not specifically describe them as “Mexican” or anything else. There were other stories that simply described a suspect as “a man” or something similarly nondescript. We of course could have simply labeled all of these as “vague,” but surely readers of the time would have made certain presumptions based on such details. Thus, in the American Southwest, a person with a Hispanic surname surely would’ve been assumed to have been of Mexican descent. At the same time, I think it’s also fair to presume that if the paper didn’t bother to describe someone beyond identifying them as a “man” or “woman” or something similar (e.g. “a plumber”), a reader would’ve presumed them to have been a “normal American” which for the mostly white readership meant White and probably Anglo-Saxon/Protestant too. Thus I made the decision to categorize all Hispanic surnames as “Mexican” and all references to specific but otherwise nondescript people simply as “Americans” in the way that a reader of the time probably would’ve understood that term. Again, like with our “Oriental” categories, these choices were based on how we imagine that readers of the early twentieth century would have read and understood these stories. There is certainly some room for disagreement there, but I think our method provides a reasonably accurate interpretation of the evidence.

Third, and similarly, stories sometimes described Europeans of various kinds, usually Parisians or Londoners, who were involved with cannabis, thus we created a broader category of “Europeans” to consolidate all such references.

Fourth, because the demographics of users would have had a different valence depending on whether they were close to home or abroad, we added qualifiers to our categories to include both “‘Mexicans’ in the U.S.” and “Mexicans in Mexico,” “‘Orientals’ in the ‘Orient,’” and “‘Orientals’ in the U.S.’” and so forth.

Fifth, because we were interested as much in the sellers of these drugs as the users, and because the sellers were often entities like patent medicine manufacturers, and because the users were sometimes animals (in the case of veterinary medicine), we decided to use the somewhat cumbersome terminology “seller/user ID” for this category and include “animals,” “patent medicine manufacturers,” and “Mexican groceries” (this final category being a key one in the “Mexican hypothesis” as presented by Richard Bonnie and Charles Whitebread).

Sixth, because there was sometimes (but rarely) more than one demographic mentioned in a story, we noted up to three different sellers/users per article. As when we were analyzing nomenclature and effects, to do this we had to create separate fields, in this case “seller/user ID I,” “seller/user ID II,” and “seller/user ID III.” Once again, while we tried to put the most prominent demographic first, this was hardly an exact science, so these should not be viewed as a strict hierarchy. If you’d like see which terms were most often paired with each other when more than one demographic was mentioned, you can do so here.

Finally, we only noted a user or seller if a story specifically cited a person, group, or entity in connection with the drug. Thus, for example, a story that simply called hashish an “oriental drug” or marijuana a “Mexican drug” was not enough to justify an entry under “seller/user ID.” Such stories would have had “none” as their “seller/user ID.” However, if a story noted something like, “marijuana is a favorite of Mexican soldiers,” or “hashish is used widely by Egyptians,” that would have been enough to justify entering the respective demographic label.

The Findings

Let’s now turn to the data, beginning with just a basic count of the groups or entities that were identified as sellers and/or users of cannabis.  Not surprisingly, given the dominance of “hashish” stories, the most commonly cited demographic category was “‘Orientals in the ‘Orient.’” But beyond that there are some quite interesting wrinkles:

First, most stories did not actually name a specific demographic as users or sellers of the drug. There were more than twice as many stories with no demographic category as there were articles mentioning the second most common demographic group (“‘Orientals’ in the ‘Orient’”). This is a very interesting finding given how much weight so many authors have given to user identity in trying to explain both anti-cannabis attitudes and laws. We’ll examine what that might mean a little more thoroughly below.

Second, and perhaps not surprisingly given the number of “hashish” and “marijuana” stories, “Orientals” and “Mexicans” were especially prominent. But it’s also noteworthy how many “‘Americans’ in the US” appeared in these stories. This clearly was not a substance exclusively linked to “the other,” though, again, what that might actually mean will require some further exploration which we’ll do below.

The vast majority of stories only included, if any, one demographic category, but here’s what the second category looked like:

And the almost nonexistent third category:

Demographics and Cannabis Nomenclature

What was the relationship between these demographic categories and the various terms used for cannabis?

Not surprisingly, hashish was overwhelmingly associated with “‘Orientals’ in the ‘Orient,’” and marijuana with “Mexicans” in both the U.S. and Mexico, though by far the biggest category of articles involved “hashish” with no specific demographic mentioned. There were also a significant number of articles citing use by “Americans” in the U.S. (in the case of “marihuana”) and “Europeans” in Europe (in the case of “hashish”). Again, returning to the questions we posed above, if the identity of the sellers/users was a key factor in concern about this drug, rather than simply its reported effects, what specifically was driving that concern? Was it that this was a substance being used by foreign people within the United States? Or was it that this was a foreign substance that was spreading to “American” users? One imperfect way we might further examine this is by comparing how these numbers differed with respect to the word “marihuana” between the English-language and Spanish-language press. This is imperfect because most of our Spanish-language stories here (59 of 75) came from San Antonio’s La Prensa which was founded by Mexican revolutionary exiles with the intention of keeping the exile community up to date on events within Mexico. Thus their coverage was decidedly slanted toward happenings south of the border. But with that important caveat in mind, we might look at the numbers anyway:

Keep in mind that the corpus had vastly more English-language newspapers available, and thus vastly more articles in English about cannabis (1,149) than in Spanish (76). Thus even though there were twice as many articles using the word “marihuana” in the English-language press, this was a much smaller percentage of the total English-language cannabis articles than the Spanish-language ones where, not surprisingly, 75 out of 76 total articles used the Spanish word “marihuana.” Given the paucity of Spanish-language articles in comparison to English, we should probably take the above percentages with a grain of salt. But they are nonetheless quite interesting in that they are relatively comparable in most of the categories save for reports of use by “‘Mexicans’ in the US” where there is a huge gap, with 35% of the English-language articles, and only 5% of the Spanish language, showing those results. This could indicate that use by Mexicans in the U.S. was more noteworthy in the English-language press because it was considered more problematic. There were also 8 Spanish-language stories with “positive” connotations versus none in English:

But, again, La Prensa was mostly focused on events in Mexico, and those “positive” stories were all advertisements for medicinal products that happened to include the word “marihuana,” so these were hardly endorsements of the drug’s intoxicant effects.

Let’s then break down these results a little further and examine how often portrayals of cannabis effects were negative or positive in relation to particular demographic groups:

There’s a lot to chew on here so let’s just move from top to bottom. First, with the term “cannabis indica” we see something somewhat expected:

This was a term used for medicinal cannabis that could have positive connotations in the abstract, but when tied to “patent medicine companies,” which were routinely accused of exploiting consumers with dangerous and addictive products, the drug was usually portrayed in a negative light.

Second, “marihuana” was overwhelmingly portrayed in a negative fashion no matter the demographic group involved (the “positive” stories were all simple advertisements for medicinal products):

This surely reflects both the tenor of the Mexican “marihuana” discourse as it spread to the U.S., and the generally negative portrayals that, as we will see below, accompanied the presence of these foreign substances in the U.S. or Europe, no matter what the identity of the users.

Third, notice how with “hashish,” “‘Orientals’ in the ‘Orient’” could be portrayed in various ways, most often negatively, but often with both positive and negative connotations, or in a neutral light, while negative portrayals were very often connected with no demographic group at all.

It seems that a link to a marginalized group was certainly not necessary for negative portrayals of this drug, while such a link could still generate some positive references.

Given that so many of the articles citing “hashish” were fiction, we might ask if there was any significant difference in the frequency of negative and positive portrayals between fiction and non-fiction stories:

It is quite interesting here that with respect to “‘Orientals’ in the ‘Orient,’” non-fiction stories appear to have been more likely to include “neutral” portrayals, or ones that included both positive and negative aspects (i.e. “both”). Straight “positive” portrayals were most likely to appear in non-fiction stories associated with no demographic group at all, but there were more than double the number of negative stories that fit the same criteria. At the same time, “‘Americans’ in the US” and “Europeans in Europe” appear to have generated a predominant number of negative non-fiction references, again suggesting that it was a kind of contagion of local populations by these foreign substances that inspired the most concern.

Let’s examine that further by simply looking at each demographic group and whether cannabis, under whatever name, was characterized as a positive, negative, or something in between:

While “‘Mexicans’ in the U.S.” is the most overwhelmingly negative category, “Mexicans in Mexico,” (of which a number of stories were drawn from the Mexican press), “‘Americans’ in the US,” and “Europeans in Europe” are not far behind. The one really nuanced category was “‘Orientals’ in the ‘Orient’” where stories were nearly as likely to be “neutral,” or both positive and negative, as they were to be simply negative. These numbers of course do not explain exactly what was happening in these stories, but they do belie the idea that the effects of cannabis were only portrayed in a negative light when associated with “the other,” whether domestically or in a foreign context.

We might then wonder if certain specific effects were more commonly reported when the drug was linked to certain demographic categories:

A couple of trends stand out here. First, for “‘Orientals’ in the ‘Orient,’” “alt-perception” was the most likely reported effect, and for “Mexicans in Mexico,” it was violence, the latter presumably reflecting the dominant stereotypes of marijuana’s effects in Mexico. On the other hand, for “‘Mexicans’ in the US,” the most likely effects were “vague.” We described effects as vague when clearly the use of the drug as an intoxicant was involved and there was a “basic effects” spin on it (i.e. “positive,” “negative,” “both,” “neutral”), but beyond that there was not a description of those effects. The most typical example of this was when an article described the arrest of someone for intoxicant cannabis, arrests by default having negative connotations in our system, but then no further explanation having been offered of those effects. Notice how “vague” references tend to be in higher proportion to others when that use is being described in either the U.S. or European context. While some of these reflect short stories on arrests in the U.S., the same being true in Europe might suggest that use of the drug in these contexts was already considered “news,” and further description was considered of secondary importance.

In short, our data seems to support the most recent research which suggests that demographic factors alone do not by themselves explain positive or negative portrayals of cannabis in the media, or at least not in any easily predictable way.

Next Post: Cannabis Geography