This blog is no longer active, you can find my new stuff at: george3d6.com, epistem.ink, ontologi.cc, and phenomenologi.cc

Vegetarianism and how the ethics of FOSS can get muddled

An analogy with veganism will be very helpful to illustrate my point. Since I assume most people reading this aren't vegan, but many of them might be advocates of using FOSS.

First of all, I should say this article comes from a deep personal bias towards FOSS, I myself am somewhat of an advocate of FOSS.

I don't refuse to use or release software under anything but a copy-left license. However, there's hardly been any closed-sourced software on my laptop in the last 4 years, barring a few necessary binary blobs here and there.

I don't freely contribute to any popular FOSS repositories, but I do try to do my part in reporting issues or making a PR every now and then if I can easily help.

In the following analogy, I'm the FOSS equivalent of the vegetarian that eats honey, eggs, goat cheese and takes a customary bite of pork/lamb at the christmas table in order to not upset grandma.

To me, this stance seemed rather reasonable. Until I started musing on the topic of vegetarianism...

On vegetarianism

I have a bone to pick with the reasoning people have for being vegetarians. There are three pillars on which a vegetarian diet rests:

The ethical arguments against killing, factory-farming or causing pain and suffering to animals.
The environmental argument, according to which, farming animals or hunting wild animals is more environmentally taxing than cultivating or gathering plant-based food and materials.
The health benefit arguments, according to which, a "correct" vegetarian (aka vegan) diet will lead to increased lifespan and healthspan.

All three pillars stand pretty well on their own, but you can poke holes in each one of them.

The health benefits argument is the easiest to argue against. Since most of the studies touting the health benefits of vegetarian diets are confounded by a host of other factors.

Ranging from vegetarian diets being practiced in communities with other health-conscious habits, to vegetarian diets making it easier to avoid "food traps" (e.g. processed meat).

Ultimately, it's easy to argue against it as a foundation for the same reason almost all other dietary advice is, metabolic science is too complex to make any general claims at the moment.

The environmental argument can be combated using the same utilitarian principles on which the other two arguments are based.

That is to say, even given that animal farming is less beneficial for the stability of our environment than farming the equivalent amount of calories and nutrients as plants, the shift in zeitgeist needed to stop or significantly reduce it is not worthwhile compared to other ways of improving the environment.

Good examples here would be investing money/energy into promoting nuclear energy, or shifting away from fixed-location animal farms to roaming herds in order to combat desertification.

In other words, there’s only so much time for changing culture and policy in a short amount of time, and replacing animal farming is hardy as impactful as other alternatives, where a shift in culture and policy might be easier to trigger.

Lastly, the ethical argument can vary depending on the framework in which it is created, but holes can be poked in most ethical arguments by virtue of them resting on the previous two pillars.

Singer's argument for vegetarianism is a classical example, easy to understand, well-aged, still the basis for much of vegetarian ethics. It avoids basing itself on the other 2 pillars of vegetarianism, except for one small slip:

The overwhelming weight of medical evidence indicates that animal flesh is not necessary for good health or longevity.

Now, this is a rather tame invocation of the veganism-for-health pillar, it seems that Singer's argument doesn't rest upon it. However, by introducing it, Singer glances over the possibility of discussing a rather important flaw of his reasoning: Would it still hold true if, hypothetically speaking, eating animals would be beneficial for human health ? Furthermore, what degree of evidence and what kind of benefits would be required to accept eating certain types of animals raised in certain types of conditions.

By glancing over this point, most ethical arguments refuse to pose a question such as:

How many years of healthy living would the consumption of y grams of {X} meat have to add to a human lifespan in order to be ethical to factory farm {X} barring another maintainable source of supply ?

There are thousands of variations on this question, all of them lead to the rather uncomfortable topic of discussing the importance of animal life and suffering compared to human life and suffering.

Other holes can be poked into ethical arguments for vegetarianism. Such as slipper-slopying them into a sort of antinatalist stance for animals, since both of the stances are terribly close in their evaluation of not-existing being preferable to a certain amount of suffering, they only vary as to what that degree of suffering is.

My point here is not to "debunk" vegetarianism. It's only to say that these pillars that it rests upon can be viewed as somewhat shaky and argued against with a great possibility of success. Each of them is insufficient in of itself to convince the vast majority of people to become vegetarians. At the very least, you can mount and argument and debate each of them without necessarily reaching a black and white conclusion.

In practice this is not possible. One cannot go up to most vegans and debate each pillar in absentia of the others, since they will simply "switch" the discussion to another pillar, forcing you to debate that particular pillar. This circling-around effect means that you’d need skills in long-term economy, climate science, philosophy of ethics and medical nutrition… and a whole lot of time on your hands, to change their mind.

Even if a proponent of veganism will probably not be able to hold a view that consistently validates all three pillars, he will be able to hold a view that allows for none of them to be fully invalid . From this viewpoint, he will combine the 3 shaky arguments and claim that the whole is more stable than the parts.

… Going even further, the vast majority of vegans will be too poorly informed to actually properly understand these three pillars. They will instead, be swayed with two additional factors:

Appeal to disgust

Most people aren't swayed into veganism after reading a well thought out ethical argument. Instead, they start considering it after seeing a shocking (and possibly fake) video portraying the factory farming of animals.

Their main drive to be vegan is the disgust response at seeing animals treated in a certain way, the arguments for why treating animals that way is wrong, are often secondary or too simplistic to hold under scrutiny.

Peer pressure

The second factor, rests on the fact that vegetarians often form a community. Around vegetarian restaurants, vegan-friendly shopping centers, cooking lessons, apps that let you find the "unethical" ingredients or testing methods in products... etc.

Most people get involved in this community to some extent, they make friends and acquaintances within it, they become socially and emotionally tied to it.

Which gives a diminishing return to the prospect of switching back to consuming animal products. After all, most people that have been vegan for 10 or 20 years, might have an entourage consisting of a rather large number of other vegans... many of which would be appalled if they stopped being a vegan.

So then the problem stops being an issue of ethics, or disgust or any other compatible stance. Instead, it becomes an issue of keeping your standing in your social circles.

In the end we are just monkeys with language, we might be convinced by an argument even if it involves sacrificing certain things, but we're hardly going to be convinced by any argument to sacrifice out standing or even membership within a community we are part of. Especially since making friends and becoming members of other communities becomes harder and harder as we age.

Aren't we supposed to talk about FOSS ?

Well, I'm getting there.

Consider whether or not you'd think of vegetarianism as a stronger position if, instead of the above, there was a single reason why people decided to become vegetarian.

That reason could be a sound argument for why animal consciousness should be treated similar to human consciousness. Or it could be a scientifically solid theory of nutrition that shows there are always superior plant based alternatives to animal products.

I think the same may be the case with free and open source software. That is to say, most of us think we use it and/or write it due to easy to “strong” reasons. But when we look inside, it's a constellation of half-baked ethics, emotions and empirical pragmatic/survivalist reasoning that guides us instead.

Let's consider three potential pillars on which the usage of FOSS usually stands:

The ethical argument. Which boils down to humans having the right to program computers and not vice-versa and to the principle of equality between humans, including access to the same knowledge if desired
The cost concerns, 0$ being better than x$.
The quality argument, that is to say, the quality of foss is often better due to the amount of people contributing to and scrutinizing the code.

I think the ethical argument is by far the strongest, and it can be pretty well summarized by the "philosophy" section on gnu.org. Or, to some extent, even by this single piece written by Stallman.

However, the ethical argument fails to contend with some real world facts. Even ignoring the economic realities of our societies, it fails to contend with the importance of distribution and ownership of hardware.

A good example of this is Facebook, which open source most of its tools, and could presumably open source most of its code without much fear. However, it's not Cassandra, or torch or Hack that makes Facebook a social media monopoly, it's their user base and the data they have about then.

An even better example of this is Google, which open sources most of their "bleeding edge" ML models via the code & papers published by Google Brain and Deepmind. Sadly enough, this does little to help anyone, since Google's ML code is designed with massive amounts of parallel hardware in mind.

The ironically named Efficient net is a great FOSS advancement in computer vision, for anyone with a few server farms of TPUs worth to spare, which boils down to about one company.

So, in a way, OSS will lead to situations where writing software (something everyone can do) losses socio-economic value and collecting data or building hardware (big companies and monopolies have a huge advantage) becomes more valuable.

The cost argument seems pretty solid, but I'm sure we've all heard the counter-argument to it. That is to say, if paying for software helps paying for less development time, the former is usually cheaper than the later.

Which is almost certainly not true, in the case of something like Oracle Db vs Mariadb and it’s almost certainly true in the case of something like IntelliJ vs Netbeans. However, most cases aren’t as clear-cut as those.

Finally, the distributed vs centralized development argument is well illustrated by the Cathedral and the Bazaar.

This argument is similar to the environmental argument we had with veganism.

It sounds "nice" on the surface but it's mostly based on good rhetoric rather than objective facts or logical deductions. I wrote a bit on my thought on the subject in my article about choosing programming tools, my point there boils down to us choosing most of our tools based on superficial values, rather than based on in-depth studies into their previous usage.

We can almost certainly say Linux is a better Kernel than Window NT in almost every aspect. But we can't hold that example, or any particular example, as fully representative of centralized vs decentralized development.

Now, again, the fact that these pillars, are imperfect, does not negate the validity of advocating FOSS, or advocating FOSS in certain scenarios.

Nor am I claiming these are the only pillars on which one could rest the merits of using & writing FOSS.

However, any argument for using FOSS that we might come up with, will undoubtedly rest upon a multi-faceted and imperfect foundation.

We must also keep in mind that whilst we bring up these more objective arguments to bare when we argue for open source software, we also usually use a disgust response, fear response and a sense of community to strengthen our ideology.

The disgust or fear response could be against the companies promoting closed source software. We might hate Windows because they killed a bunch of competitors through unfair practices. We might hate Google because of illegal data collection and patent trolling. We might hate IBM because it bought Redhat or even because it sold hardware that helped propagate the holocaust.

However, these are all appeals to emotion rather than appeals to reason. We can't confuse the ethics of specific creators with the efficacy of the software.

Otherwise, we could apply the same logic to FOSS. Claiming that the opinions we hold about someone like Richard Stallman or Linus Torvalds as a person, are directly linked to whether or not one should use gcc or linux.

Lastly, there's the issue of community. And here, I think, is where the ethical gap between FOSS advocates & users and closed source software advocates turns into a social gap.

Let me once again give myself as an example. I earn my living by creating an open source machine learning package using FOSS and by building data-centric systems using FOSS. So, in many ways, I'm in direct competition with closed software producers.

If the closed source tools provided by makers like Google or AWS for data analysis would be a good bang for the buck, I'd be out of a job. (happily enough, they are utter garbage in terms of performance vs cost)

If a closed source machine learning solution (which can obtain much higher profit margins) is equal or preferable to an open source one, I'm out of a job.

Going even further, most of the developers I know work with FOSS. So, if closed source becomes the dominant tooling paradigm once again, not only am I out of a job, I'm out of a network to help me find work.

Even more, say I have a change of heart and want to get deeper into the proprietary software world. The whole "standing" I might have in my current social network doesn't apply there.

Having one or two interesting projects on Github goes a long way in the space I work in. But would be completely irrelevant if I want to work for, say, DARPA, or IBM.

It seems to me that, in many ways, being an advocate of FOSS can easily lead you into the same pitfalls as something like veganism... or being part of a cult, for that matter.

Is that to say I now suddenly dislike FOSS or that I will use it less ? No, not at all. Nor am I saying that you should.

But I do believe that we should be a bit more critical when discussing the ethical merits of FOSS. That is to say, we should admit that we don't necessarily have a fully solid foundation to hold this ideology. More importantly, we should admit that we might rationalize our preference of FOSS in ethical reasons, when it actually comes, at least partially, from our emotions and our social constraints.

If you enjoyed this article you may also like:

Published on: 2019-10-16

Reactions: