I research how people learn new skills using instructions. As fields, I selected working with people who learn programming and people who (re) learn cooking (some of them do so as they got a diagnosed with a condition that demands a new diet). Both fields work with very different materials. Cooking is (among other things) concerned with taste, smell, texture; programming with abstract structures, text and logic. I selected the fields because, they have strong cultures of using instructions, despite of their different materials and concerns. Both fields have a lot of books as well as an exchange of shorter, participant created instructions as videos or texts.

Several times people asked me to consider if or how recipes are means of control and prescription. They mentioned different sources that the discussion of this issue should include:

• Seeing instructions in a family of resembling concepts: “norms, directives, regulations, laws, commands, order, rules, standards, plans, programs, budgets, maps, manuals”, a list cited from Garfinkel's “Ethnomethodology’s program” ¹ Laws and commands obviously connected to an expression of authority whereas maps and plans might be less associated with them.
• The debate of situated cognition vs. symbolic cognition in the 1990s. The most famous text here is probably “Plans and Situated Actions” by Lucy Suchman ². The conflict proposed and worked on is that human actions can be understood as situated and strongly driven by context but that cognitive science assumes actions are executed based on previously formed plans.
• Wittgenstein on Rule-Following, specifically §201 in Philosophical Investigations ³: “This was our paradox: no course of action could be determined by a rule, because any course of action can be made out to accord with the rule”. If one assumes that the prototypical mode of engaging with instructions is following them like written plans for human actions that are then executed, these instructions function as a rule one should behave accordingly to and thus the paradox would also apply to the following of instructions.
• Bourdieu's argument that self-directed learning (e.g. from instructions) does not lead to power that can be used by the learner. In Distinction, Bourdieu discusses learning from books and focuses on learning the “right taste” concerning movies, music and artworks ⁴.

The first three relate to my research when seeing instructions as rules to follow, as they all suggest that instructions strongly structure actions. Bourdieu's observation I look at separately, as it is not about the use of instructions itself, but the social role of knowledge learned by instructions.

Instructions as strongly configuring actions in the situations of use

The ideas of people executing abstract plans and of instructions as externalized and reified copies of these plans suggests that instructions strongly structure actions.

Comparing instructions to rules, laws or standards does not assume an innate equivalence to assumed cognitive instructions for behavior. But rules, laws or standards do strongly configure actions by the implied thread of punishment when not followed.

I had some trouble relating to the question of power and instructions. This was also somewhat to be expected: My main focus is how participants interact with instructions when learning, and thus I selected fields where I could focus on this. I did not pick fields where the near-term goal of participants was to pass a test or to show the correct taste (like in Bourdieu's examples), nor did I choose fields where there is a strong focus on following rules correctly (which would concern Wittgenstein's paradox). I rather did choose fields where activities appeared to be strongly participant-driven, where people would say “I would like to learn Javascript to build a web application” or “I would like to be able to bake a gluten-free chocolate cake”.

It is hard to draw a boundary, though, as people who are pushed into formal structures, for example with an instruction that “tells the right steps” will also learn how to deal with these prescriptions in a situated way, so I would not claim that learning (partly) to pass a test or satisfy a formal requirement excludes skill. Neither would I claim that learning that participants see as self-directed exist outside of power structures. But in my selection of fields to study I wanted to focus on the aspects of learning that would show in situations in which participants perceive themselves to be under control and in which they need to evaluate their own actions as successful or not.

Interestingly, the idea of “using instructions correctly” was brought up by participants themselves when they talked about cooking or baking. This came up less in the interviews but when talking with people informally about what I research, along the lines of “You research recipes? I am terrible at following recipes”. When I asked what they mean by that, they explained that they did not follow them “exactly”, which means they only roughly guessed amounts or replaced ingredients with what they had at home or what they liked: They had no shallots, so they used onions. Some participants used recipes as starting point for remixing them, taking the cake from one recipe, the icing from another. This all can be seen as “not following the instruction“. The assumption seemed to be that to qualify for participation in my research, diversions from recipes are problematic. Actually, I was quite curious about the diversions.

I can only speculate why people assumed this. One hypothesis is that they just build upon the logic that if I research recipes, people assume two prototypical ways: Structuring their whole work according the recipe if anyhow possible or not using a recipe at all. Not using a recipe at all was something that the research question “how do people use recipes” could not be researched with, and the less people structured actions according to a recipe, the less valuable they assumed their use of recipes might be.

Another hypothesis is that recipes are often perceived as a stand-in or replacement for an experienced person and thus an authority. If instructions are written by authorities then “not following” the instruction deliberately is deliberately violating the authority. However, this hypothesis is contradicted by the joy that many people seem to take in sharing their diversions from recipes. Below many recipes shared on websites are comment sections and a lot of the comments are usually what people changed about the recipe, commenting that they think it improves with some cinnamon, still tastes great with less eggs or that the milk can replaced by a vegan alternative – and if there is a reaction by the authors of the recipes, they seem to appreciate these contributions. In my interviews people also directly commented on them using instructions very much not like instructions-to-be-followed: They told about reading only parts they needed to answer a particular question when programming or mixing different recipes to create new variants of an existing dish.

So far, I do not see instructions as necessarily strongly configuring actions in my research. Participants use instructions based on their situational sensemaking and based on the resources they can access in the situation. Indeed, building upon existing instructions, as in the case of recipe variants, can be a source of pride and a contribution to the community. However, this is partly because I deliberately choose fields where there are no formal authorities that would punish diverging from the instructions.

Instructions and their social effects outside of the situation of use

In Distinction, Bourdieu shows that learning by books does not create social capital. This is contrasted with the social capital that is generated by implicitly learning the skills in one's social context.

However, Bourdieu focuses here the education of taste art, music, interior and food. Even though food and tasting take up a large part of my research in the field of cooking and baking, I focus on the food being judged by the cook and maybe their family. This is different from the perspective that Bourdieu takes, that is, a judgment that generates (or not) social capital. Obviously, the food that is cooked can play an important part in showing that one has “taste” but this is not focus of my research on instruction use. Nevertheless the use of instructions itself has social implications on its own that I will outline.

Instructions as expression of rationality

In Orr’s ethnography “Talking about Machines” ⁵, instructions are tools of expression both of managerial power and of repair technician's competence in repairing photocopy machines. The management says that technical problems can be resolve by actually mechanically following the steps in the instructions. They present them as universal and algorithmic and thus implicitly as a tool of deskilling: Anyone could repair a photocopy machine by just following the instructions (Orr 1996, p. 104). However, Orr shows the tacit skills that are needed for repairing machines efficiently. The technicians trust the instructions only somewhat. But they do not reject them. They are not only a way to suggest deskilling by management to workers but also a way for workers to show themselves and clients that their work is orderly and controlled (Orr 1996, p. 110)

The use of instructions in “Talking about Machines” resembles the use of math in Lave's “Cognition in Practice” ⁶ where using abstract math with the abstract principles learned in school is an expression of rationality (Lave 1988:158). The work done with math when not used for expressing rationality was far more situated and did barely involve formal abstractions.

Instructions as configuring people and infrastructure

All use of instructions needs to assume existing skills in interpretation and a specific infrastructure. This means that people need to know how to interpret “bring to a boil” and “add some salt to the water", and they also need a stove on which they can boil water, access to ingredients, a pot etc.

The skills and the infrastructure can be configured by instructions to provide a more favorable context for companies: In Germany, one of the best known brand for cook books is the same brand that is also well known for baking ingredients – Dr. Oetker. Their recipes will usually refer to at least one branded ingredient. If you do not know how to replace them with generic counterparts, it is a safe way to just buy the Dr. Oetker product that is directly referred to in the recipe ⁷. This is not deterministic: Many people will e.g. not use the branded baking powder but replace it with a generic variant that is cheaper. But one needs to know that they are practically equivalent and thus it favors the company.

Thus, instructions are not one-sided self-empowerment of their users: They dissolve some ties (e.g. to a mentor or master) and create others (e.g. to the companies creating these instructions)

Usefulness of instructions being contested

Just as the managers in Orr's “Talking about Machines” insist that work can and should be done based on the provided instructions, some people also might push in the other direction, emphasizing that an activity can't be learned or helped by an instruction.

In his Ethnography “Body & Soul” Loïc Wacquant describes learning boxing from his mentor and trainer. The trainer is insisting on the impossibility to learn boxing from books. This is a brief excerpt, starting by Waquant asking:

“So you can’t learn anything about boxing in books then?”

“No, you cain’t.”

“But why not?”

In a tone irritated by my insistence, as if all of this went so much without saying that it was useless for him to repeat himself: “You just can’t! Period. You can’t. In a book, everything’s standin’ still. They don’ show you what’s happenin’ in d’ring. Tha’s not boxin’all that stuff, Louie. You can’t, tha’s all."

(Wacquant 2006:101)

While it does not necessarily need to be the motivation, the trainer's role depends on teaching, a role possibly threatened by instructions: If it would be possible to learn boxing (at least somewhat) from books, his authority would be diminished.

Learning from instruction can also be accepted, but seen as of low status: Participants in Lange's research of kids' media use on the video sharing platform YouTube⁸ do not reject instructions for learning. But instructions are not the most prestigious way of learning. Better than learning from instructions is learning by trial and error, which is presented as being more true and independent (Lange 2014:192, 195, 199). This is different from Bourdieu's observation that learning from books does not lead to behavior that can be converted to social capital, since in Lange's research participants discuss the source of learning directly and while Bourdieu suggests that the socially valuable learning happens by early exposure and by learning from others, Lange's participants emphasize their independence by actually not having ties to others.

While in these two examples the focus is on how instructions are seen by participants in a community of practitioners, an instruction user might themselves decide in their situation of use that some practices can't be learned by instruction. In his reenactment of instructions for gold- and silversmithing⁹, Thijs Hagendijk concludes that the tasks can not be done based on the guidebook he used alone as "effective use of the Guidebook depended on complementary hands-on education of master craftsmen, which suggests that the Guidebook was far from a DIY crash course and illustrates that the textual transmission of craft knowledge depended upon, rather than threatened, established routes to craft learning, such as apprenticeships” (Hagendijk 2018)

However, Hagendijk also suggests that learning from Workshop masters had its problems and gaps too: “The Guidebook thus seems to be a clever solution to overcoming the challenges posed by hands-on education, such as divided attention from the master and limitations brought about by specializations, prioritized work or plain didactic clumsiness.”

In my research, participants did not seem to worry much about the question of what could be learned from instructions -- they seemed to use them where it made sense for them. Thus, I also did not encounter situations where people rejected instructions as useless. It is plausible to assume that they were competent in their use of instructions so they could decide what made sense for them. They could detail, however, why they preferred certain ways of instruction and what it helped them. This was not necessarily what was most similar to observing a master in their practice: Videos that could be slowed or sped up or series of comparisons (e.g. what food looks like when baked for different times) helped participants to make sense of their actions despite not being like observing an every work process.

Conclusion

Instructions are often associated with other genres that have obvious relations to control and power, like Garfinkel's associations of ”...norms, directives, regulations, laws, commands, order, rules, standards…” (Garfinkel und Rawls 2002:199). These associations match the idea that instructions prescribe actions and that any deviation is either an error of the instruction's user or the instruction's author as well as symbolic views of human cognition as the internal creation of abstract plans that are executed. While instructions can be used to enforce actions and behavior, this is not an essential quality of instructions itself: Participants in my research did use instructions creatively and based on their perceived needs in the situation they acted in, more similar to how people would use a map (A genre that Garfinkel mentions, too).

Similarly, my participants did not see the use of instructions as superior or inferior to other means of developing skills. Nevertheless, instructions are part of networks that also aim to configure behaviors and assumptions like the use of certain ingredients or the authoritative way to do an activity.

The concepts of both instructions as strongly structuring and thus controlling and the view of knowledge learned by instruction as the inferior knowledge do propose an either/or distinction between use of instructions or other ways of learning – Even though Wittgenstein as well as the situated knowledge research movement are quick to deconstruct it again, focusing on what instructions can't do.

However, if instruction use is itself a skillful activity, there might be no either/or distinction between use of instructions and other means of learning. If an instruction is useful for a learner might not hinge on the instruction fulfilling the promise of being like another form of instruction (and failing by this yardstick) but on how it is usefully different than other ways of learning.

---

Notes

• Related book: Daston, Lorraine. 2022. Rules: A Short History of What We Live By. Princeton: Princeton Univers. Press.
• A lot of instructions are very light on directing their user in an obvious way. While there are instructions in the style of “You should do this, then that” (particularly in operating instructions for devices), others might be written in a distanced style of a protocol, rarely hinting at a connection with the user’s actions at all. For example, this tutorial on game programming uses an inclusive “we create…” style and consists mainly of discussing what the example code does.