All the ideas in our mind that are not simple are complex. These complex ideas come in four basic varieties: modes, substances, relations, and abstract generals. Modes are ideas that do not include any notion of self-subsistence, in particular, qualities, numbers, and abstract concepts; qualities depend for their existence on substances, whereas numbers and abstract concepts do not have any archetypes out in the world, but exist only as ideas. There are two types of ideas of mode: Simple modes are created by taking a single simple idea and either repeating it or varying it (examples include "dozen," "infinity," "oval," and "space"). Mixed modes are combinations of simple ideas of different kinds (examples include "murder", "obligation", and "beauty"). In contrast to modes, substances are either self-subsisting things (e.g. a man or a sheep) or collections of self-subsisting things (e.g. an army of men or a flock of sheep). Relations are simply relational concepts, such as "father," "bigger," and "morally good." Abstract generals are not treated until Book IV. Complex ideas are created through three methods. First, simple ideas can be glued together through combination, either by taking stock of simple ideas that come into the mind together naturally though sensation (for example, gluing together yellow, long, wheels, loud, etc. into "school bus") or else by mixing and matching simple ideas in the imagination (for example, to create the idea of a mythical creature). Complex ideas can also arise through a comparison of simple ideas, in which we take two or more simple ideas and observe the similarities and differences. This method results in the complex ideas of relations. Finally, there is abstraction, in which the mind separates ideas previously joined by the mind.
Chapters xiii-xx analyze our ideas of simple modes, focusing in turn on the ideas of space, duration, number, infinity, pleasure and pain, and powers. Examples of ideas of space include "space," "place," and "inch," and are produced by considering two ideas of color or texture and noticing the distance between them. We form ideas of duration, such as "time," "year," minute," "eternity," by noticing that we have a train of ideas and that this succession has distances between its parts. The idea of number is produced by repeating the simple idea of unity. Realizing that there is no end to the process that gave us the idea of numbers produces the idea of infinity. Ideas of pleasure and pain, such as "good," "love," and "sorrow," which are produced in reference to our simple ideas of pleasure and pain. Finally, we get ideas of powers, such as the ability to cause things to melt or the ability to be melted, by perceiving changes in our ideas and noticing that these changes happen in regular patterns.
Chapter XXII examines mixed modes. Mixed modes, Locke tells us, are created simply for purposes of communication. We glue together certain ideas by giving them a collective name if and only if collectively they will prove useful in discourse. So, for instance, we decided to glue together the ideas of murder and father into "patricide," but it never proved as useful to glue together the ideas of murder and son, or murder and neighbor. To strengthen his claim that mixed modes are invented for reasons of convention, Locke points out that often one language will have a word for a concept that does not exist in another culture. He also points out that languages constantly change, discarding and creating new mixed modes as our communicative needs alter.
Locke's application of the categories "substance" and "mode" is rather unique in the history of philosophy. Both Aristotle and Descartes agreed with Locke that the distinguishing characteristic is self-subsistence. However, for them, only actual objects were self-subsistent; they would not have included collections of objects as substances. It is not entirely clear why Locke feels the need to classify collections as substances, since collections do not really have any self-subsistence out in the world in the way that single objects do. He probably could as easily called collections mixed modes rather than substances and account for their origin in the same way that he accounts for the origin of concepts like "dozen."
Aristotle and Descartes also limited the term "mode" to those things that depend on substances for their existence in a very literal way. Qualities were modes for him; abstract concepts were not. It is clear, though, why Locke felt justified in enlarging the scope of "mode." A mode, he felt, is not just something that is physically dependent on substances; it is also ontologically dependent on substances. We individuate modes in terms of the substances they depend on. While concepts like "murder," "gratitude," and "theft" do not physically exist in substances, they do depend on substances for their existence as ideas. We get these ideas by considering the relations and connections between our ideas of substances.
Now that Locke feels he has demonstrated where knowledge does not come from (i.e. innate principles or ideas), he sets out to show where it does, in fact, come from. This project will consume the rest of the Essay. The picture, on its surface, is exceedingly simple. Knowledge is built up from ideas (the operation by which this occurs is discussed in Book IV). Ideas come in two basic types: simple and complex. Complex ideas are built from simple ideas. All knowledge, therefore, traces back to simple ideas, and simple ideas come exclusively through experience. Book II chapters i-vii are all about the origin and nature of these simple ideas. There are only two ways that a simple idea can find its way into a human mind: through sensation, or by reflection. In sensation the mind turns outward to the world and receives ideas through the faculties of sight, hearing, touch, smell, and taste. In reflection the mind turns toward its own operations, receiving such as ideas as "thinking," "willing," "believing," "doubting." In either case, the process is completely passive. Locke breaks simple ideas down into four categories, each of which receives its own chapter. Chapter iii discusses the ideas we receive from a single sense, such as from sight or touch. The idea of blue and of the sound of a trumpet would be examples of ideas from this category. The idea of solidity, which receives its own chapter (iv), would be another. Chapter v looks at those ideas that get into the mind through more than one sense. Shape and size, for instance, are ideas that arise both from our sense of sight and from our sense of touch. Ideas which come into the mind through reflection are the topic of chapter vi, and chapter vii focuses on those ideas which are the product of both sensation and reflection. As examples of this last type of idea, Locke uses the ideas of unity, existence, pleasure, pain, and substance.
In introducing the notion of simple ideas, Locke claims that we can break all of our experiences down into their fundamental parts. If we see a cat, for instance, we can break that sensation down into blackness, softness, shininess, a certain size, a certain shape, etc. Fundamental bits, those that are "uncompounded, without parts," and cannot be broken down any further, are the simple ideas. On first blush, this definition of simple ideas seems plausible. Certainly, our experiences of the world can be analyzed down into their component parts. However, a little prodding leaves the definition looking less tenable.
Take a solid blue wall, for instance. Surely looking at this wall would yield a single simple idea. It is a prime example of something uniform and uncompounded. Consider, however, the shadows that would inevitably be cast across the wall, as well as the other minute variations in shade that would inevitably be present. Now it is not so clear whether the wall yields a single simple idea or many. One has to wonder whether there really can be an end to this analysis of experience down into component parts, whether there are any fundamental parts that cannot be broken down any further.
As an even more unsettling example, take the taste of wine. To people with unsophisticated palates, this is an uncompounded idea, but other people sense many components in a single sip of good wine. To them, this idea is complex. Locke certainly would not want simplicity to be relative, though. He wants the same experiences to give rise to the same simple ideas in everyone. This criterion for simplicity, then, seems to fail.
Luckily, Locke also puts forward two other candidates as criteria for simplicity, both of which seem more plausible than the first. One criterion is definitional: A simple idea is one that cannot be defined. For example, though we all know what blue looks, no one could give a definition of it, so it qualifies as a simple idea. This seems to hold very well in the case of colors, sounds, tastes, pain, thought, etc.--that is, everything that involves phenomenal experience. It is not as obviously true in the case of existence, unity, solidity and the like.
Locke ventures his last criterion for simplicity much later in the book. In Book III, chapter iv, section 11, Locke claims that simple ideas are those that cannot conceivably get into the mind in any way other than by experience. (In other words, there is no way dream them up or to derive them from someone else's description.) To illustrate, he tells the story of a man who has never eaten a pineapple but wants to know what one tastes like. No matter how much this man reads about the taste of a pineapple, or has a friend describe the sensation in all of its detail, this man will never know what a pineapple tastes like until he eats one. (Contrast this to a man who wants to know what a horse is. Even if he has never seen a horse, he can get an excellent idea of one by reading about them.) Again, though, this criterion seems more applicable to ideas of phenomenal experience than to ideas that do not involve phenomenal experience. Is it really so impossible to get an idea of unity without directly experiencing unity?