Some words mean many things. How do we memorize them? How should we read and remember wordplay? How should we search dictionaries when we encounter new words? These are valid questions without equally valid answers. I will try to address them here.
The question that leads to this post
I month ago one of my students wrote me this interesting letter:
Hello! How are you? I am wondering how should I memorize a single word with multiple and different definitions? Should I make a mental image that includes the two definitions? Or should I create a mental image for each definition? Thank you so much!
I took some time to think about it. Here are my answers. The easiest answer is: create a mental image for each definition. There is more than one answer since the issue is very different in various contexts.
Since I am also an engineer I will start by addressing a simpler aspect of programming. If you do not know programming simply go to the next heading.
It is called polymorphism. A function with one name can have multiple implementations. The compiler chooses the right implementations based on the inputs. How do we remember this?
Basically, if the code is good we do not have to. It is a requirement of good programming practice to have a name that is descriptive. This means that if one name is used to do different operations, these operations will have the same function.
Consider an operator like multiplication. We do not have to remember if we multiply scalars, vectors, or matrices. But there is a trick: not all operators are commutative. A*B = B’ * A’…. So when we remember that an operator is used, we should probably remember it with details.
In a similar way, when multiple classes inherit from some virtual base class, the name of the class that inherits should contain the reason for a specific implementation. We do not usually refer to the class without mentioning its specific individuality, but at the same time preserve the name and the properties of the base class. We can even address the base class when we want to.
Why should our mental programming be very different from the way we program computers?
In many memory structures we use in our mnemonics, especially the memory structures used to manipulate knowledge, there is a concept of linking. We remember the basic concept from which we inherit, and then branch out to specific implementations. For each specific implementation, we provide a separate branch in our memory structure. When we build a dictionary (fancy word ontology) we can add the extra branch for the specific meaning.
Say we start with the word “peace”. Then we get “Pax Romana” which a piece guarded by a big sword. We probably have the “peace movement” which is basically a group of hippies with loudspeakers. “Peace and tranquility”, possibly a meadow? The “inner peace” can be visualized by a smiling Buddha. In each of these visualizations, it makes sense to keep a small reference to the word peace, like a detail, say an emblem of the pigeon’s foot.
In a sense, we do not want to remember the word itself, but we want to visualize the bigram of the word and its context. In a way, this is just the opposite of polymorphism in programming, but very similar to the way we handle inheritance. We have a visualization that is complete but also references its elements as a sort of detail.
Quite often when learning languages we use mental palaces. A specific group of buildings usually corresponds to a specific context, like study, transportation, profession, shopping. The context in which the word appears creates a bigram implicitly.
When we use the word “bar” it will appear several times, possibly in a different form. When we think of engineering we will remember a long rod. In the area addressing nightlife, it will have bottles and barmen. In the context of politics, we will visualize it as a fence. And in a legal sense, we can imagine a group of people in black and white uniforms seating near a long table. Then we will want a reference to the basic meaning of bar, so we will probably add the detail of a long rod somewhere in the visualization. Furthermore, when we address the word barrier, we can reuse the same rod with some extra details like support.
By the way, notice that “bar” is also a verb. We will return to this soon.
Language learning is a formidable task. Memorizing the word is just one step before using it automatically. And then we will read the language and reuse the same visualization this time for retention or speedreading. It makes sense to invest some time in creating the best damn visualization with all the details you need, as you are likely to reuse it.
Looking up in a dictionary
When we check a word in a dictionary we can often see about a dozen different meanings, especially for the slang. In my eyes, linguists often look deeply disturbed people (OCD), since nobody with a healthy psyche should want to remember all of these meanings. A reasonable person will go through all the meanings and select the meaning he needs with disambiguation. He may want to remember one extra meaning for fun.
What if a word can be both a noun and a verb? Usually, this does not change the meaning of the word. It influences only the way a sentence is built. When we generate PAO markers, we can easily encode a noun into the person or into the action, and we can encode a verb into the object. Unless we want to memorize every word of our text, we can build different sentences with the same meaning and it’s perfectly all right.
All of the reasons we described earlier will tend to fail in poetry. As poets, we need to preserve not just the meaning but also the rhythmic structure of the text. Quite often this is done by dual coding. We remember the meaning as a complex memory structure, but we learn the text the old way, singing it. Then if we forget what comes next we pull the word from the memory structure. It is very hard to pull the wrong words and preserve a perfect rhythm.
To be honest, bigrams are not very comfortable for memorization. We usually memorize trigrams of person-action-object. This adds a new level of complexity.
If the word is used very often, we need to visualize it in three ways: as a person, as an object, and as an action. Then, if we want to create the translation we can use other corners of the same room for example to visualize each syllable of the word or maybe something else.
We can add a deeper level of disambiguation if a word can have more specific meanings. For example, the word “bar” as a rod can be interpreted as a stripe of light: “bars of sunlight shafting through the broken windows”.
Honestly, dealing with trigrams properly tends to be fairly complex. You may use a bigram of person and object with a “do not care” connection between them, e.g. person holding tightly the object.
Knowing all of these difficulties, when we write it is a good practice to add words and details that make disambiguation simple.
Quite often authors create wordplays: things that look and sound intentionally ambiguously. If so, the reader has to create several separate visualizations per meaning. This will definitely slow him down and cut his speedreading flow. Do you want this to happen? Do you want the reader to reread the same text with different intentions? Is the reader prepared to do that, or will he miss your wordplay?
We do not need to speedread all of the texts we read. Sometimes slow reading is just fine. Choose your path wisely.