I am living the dream job of a linguist right now. I spend my days deep in the manipulation and understanding of the context behind words. This morning I started to ask myself a question that should be simple but is not. What actually counts as a word?
The word “word” is strange because it is self-referential. It can refer to itself and to the entire category it belongs to at the same time. That creates a small philosophical loop. The concept of a word takes on various meanings depending on how you use it. You can “spread the word,” which means tell people. You can make a promise as in “you have my word.” You can give the word, a signal on when to begin.
Philosophically, a word is a symbol that represents an idea or reality. A word is essentially an idea encoded into sound or text so another mind can understand it. This is the business that I am in right now and it is a fascinating place.
What is really throwing me for a loop is that a word is a chunk of language that carries meaning and behaves as a unit in speech or grammar. But across all languages, the edges of what counts as a word are fuzzy. There is no clean boundary. Even more perplexing is how we as humans process words. Read the following.
“Aoccdrnig to rscheearch at Cmabrigde Uinervtisy, it deosn’t mttaer in waht oredr the ltteers in a wrod are, the olny iprmoetnt tihng is taht the frist and lsat ltteer be in the rghit pclae.”
Why could you process this? Your brain recognizes the pattern and fills in the correct word automatically. Reading is actually a rapid prediction and correction system. Your brain predicts upcoming words. Your eyes scan the text. Your brain checks if the prediction is correct. When you read, scan a room, watch a movie, whatever you are doing, you are actually just taking quick snapshots and predicting what will come next. These are called saccades. Saccades are the quick, tiny jumps your eyes make when they move from one point to another. They happen constantly when you read, look around a room, or scan a scene. Your brain is constantly building reality from fragments of visual snapshots.
“Your brain is constantly building reality from fragments of visual snapshots.”
Even though computer scientists are doing some amazing things right now, they have a long way to go to even get close to the marvelous capability of the human mind. I work with language models every day. I see what they can do. But they do not predict and correct the way we do. They do not build meaning from fragments the way a human brain does in real time.
This morning I asked what counts as a word. I still do not have a clean answer. But I know this. Words are not fixed objects. They are living tools we use to encode ideas and pass them to another mind. The edges are fuzzy because language itself is fuzzy. It bends. It adapts. It predicts and corrects.
The next time you read something, pay attention to how your eyes move. Notice the jumps. Notice how your brain fills in the gaps before you even finish the sentence. That is not a bug. That is the system working exactly as it should.
