Is a hotdog a sandwich?

The question is interesting because we don’t call a hotdog a sandwich, but it meets many of the salient criteria for being a sandwich.

When we think about a general sandwich, we typically think about its components and structure. A hotdog mostly has the right components and structure to be a sandwich, and yet it is usually not considered a sandwich.

Lumping and splitting

When scholars discuss the definitions of technical terms in their field, there are lumpers and splitters. These are general names for the two opposing camps that appear in any such discussion. The former favor a more general definition, “lumping” in edge cases, while the latter favor a more specific definition, “splitting” one category into multiple.

In our example, a lumper would say that a hotdog is a sandwich and a splitter would say it is not.

In scholarly discourse, the emphasis is often on what definition is most useful for describing the subject matter rather than trying to identify the “true meaning” of a word. For example, biologists often disagree about when to classify populations as different species or the same species. This has less to do with “correctly applying the species concept” (which is manmade) and more to do with why we do taxonomy in the first place.

Context-free definitions

In math and closely related fields, a definition is essentially a form of abbreviation. It’s a statement that we’re going to use a term or symbol to stand in for a longer, explicit description of what we’re talking about.

For example, a circle in geometry is defined as a set of points that are equidistant from a center point. In any theorem of geometry about circles, we can replace the word “circle” with “set of points equidistant from a center point” without changing the meaning of the theorem. It only makes it harder to read for humans.

This substitutability makes mathematical definitions context-free. All the information in the term is contained within the definition. A term defined this way has only a denotation and no connotation. In other words, there is no more to how the word is used beyond what its definition says.

Non-context-free definitions

Everyday words don’t and can’t have context-free definitions. They have connotative meaning, which makes them sensitive to the context in which they’re being used.

This is mostly an advantage. The richness and effectiveness of natural language is a result of being able to communicate extremely nuanced concepts by relying on context. Words alone, as a simple encoding, just can’t carry enough information fast enough. Constantly speaking in hyper-precise language is extremely inefficient.

For example, virtually any path that takes you back to where you started might be described as going “in a circle.” The exact circumstances in which it would make sense to say this, however, are difficult if not impossible to describe (people will disagree). It has to do with the contextual details like how the path was traversed, over how much time, and why.

One result of this is a counterintuitive dependence of concepts themselves on context.

Naïve realism on kinds

I’ve written in the past about naïve realism (aka direct realism), which in short is the view that reality is just as we perceive it. This is called naïve not as a sleight, but rather because it is the most obvious immediate interpretation of reality. As a result of its obviousness, it is often the default position for people to hold prior to having some reason to question it.

One aspect of naïve realism is the belief that human concepts are, by and large, natural kinds. That is, we categorize things not by making up categories, but by recognizing the categories that exist within nature.

While it is possible to hold that everything or nothing is a natural kind, there are some examples of things that typically would or would not be considered natural kinds. For example, many people would accept that basic scientific concepts like “organism” are natural kinds and most would exclude highly ad-hoc concepts like “blue objects that are currently on my desk.”

In its naiveté, naïve realism doesn’t necessarily distinguish human inventions like sandwiches. Concepts that exist for us in the present all feel about the same, which is to say they feel equally timeless.

If someone in Ancient Rome ate lettuce and meat between two slices of bread, would that have been a sandwich? It feels like the answer should be yes, because the way we identify sandwiches in the modern day is just by their components and structure. However, the answer should technically be no, because the concept of a sandwich didn’t exist at that time and so this hypothetical food only resembled a sandwich.

Now, that’s an example where the food in question would unambiguously be called a sandwich if it were constructed by a modern cook. But what about the humble hotdog?

Is a hotdog a sandwich?

Naïve realism runs into a problem with this question because it yields contradictory intuitions about what hotdogs and sandwiches are.

On one hand, the definition of “sandwich” appears to be a context-free definition of a timeless concept based solely on ingredients and how they’re arranged, and hotdogs meet that definition for many people. On the other hand, “hotdog” is conceptually not a subset of “sandwich.”

So is a hotdog a sandwich? The question can consistently be answered either “yes” or “no” so it’s really undecidable, but in my experience most people say no.

Where people go wrong (except to the extent that they’re just having fun with it) is in trying to contrive a more precise definition which successfully excludes hotdogs while including everything else that they would consider a genuine sandwich. It’s not that this is impossible to do, but rather it’s the wrong approach to actually address the problem.

The real explanation is simple: sandwiches are defined contextually, and so having the right ingredients in the right order is neither necessary nor sufficient for making something a sandwich.

In a way, what makes something a sandwich is really a Kripke-esque causal chain going back to the invention and naming of the first sandwich.

---

Photo by Veronika Benz