The most straightforward way of representing the meaning of a word using other words is the type of definition found in dictionaries \u2014 as we have done repeatedly throughout this book. A typical definition of the word tree<\/em> (compiled from a number of well-known dictionaries) would be the following:<\/p>\n tree<\/strong>, n \/t\u0279i\/ a long-lived plant with a single thick wooden stem that grows to a great height, with branches growing from its upper part, and leaves<\/p>\n<\/blockquote>\n Such definitions are attempts to summarize in simple terms the knowledge that speakers have about the referents of the word. As such, they rely on the knowledge that speakers have about the referents of the words used in the definition \u2014 the definition of tree<\/em> relies heavily on the words plant<\/em>, stem<\/em>, branch<\/em>, and leaf<\/em>, among others. Of course, in order for the definition to capture what we know about the potential referents of the word tree<\/em>, these words would also have to be defined. Typical definitions would be the following:<\/p>\n plant<\/strong>, n \/pl\u00e6nt\/ a living thing that grows in the earth, on surfaces or other plants, or in water, usually has roots, a stem and leaves, and produces seeds<\/p>\n stem<\/strong>, n \/st\u025bm\/ a central, tube-shaped part of a plant that grows above the ground and to which branches or leaves are attached<\/p>\n branch<\/strong>, n \/b\u0279\u00e6nt\u0361\u0283\/ a part of a tree that grows from its stem and has leaves attached to it<\/p>\n leaf<\/strong>, n \/lif\/ a flat part of a plant that is attached to a stem or branch<\/p>\n<\/blockquote>\n You can see the problem of circularity: the definition of tree<\/em> uses the word branch<\/em>, and the word branch<\/em> uses the word tree<\/em>. Likewise, the definition of the word stem<\/em> uses the word leaf<\/em> and the definition of leaf<\/em> uses the word stem<\/em>. In each case, you already have to know the meaning of one of the words to understand the definition of the other.<\/p>\n You can also see the problem of abstractness: The word tree<\/em> is defined using the word plant<\/em>, the word plant<\/em> is defined using the word thing<\/em>. How is thing<\/em> defined? A typical definition is the following, which uses the word object<\/em>:<\/p>\n thing<\/strong>, n. \/\u03b8\u026a\u014b\/ an object that the speaker cannot or does not want to name<\/p>\n<\/blockquote>\n A typical definition of object<\/em> is the following:<\/p>\n object<\/strong>, n. \/\u02c8\u0251b.d\u0361\u0292\u025bkt\/ something with a stable form that can be perceived by sight and touch and that is not alive<\/p>\n<\/blockquote>\n The word thing<\/em> is obviously not used in the dictionary definitions in the way that it is defined \u2014 first, it is not used because the lexicographer does not want to name it, as it is already named at the beginning of the definitions, second, it does not mean \u2018object\u2019, as objects are defined as \u2018not alive\u2019, so a \u2018living thing\u2019 would be a contradiction.<\/p>\n Instead, thing<\/em> is used in the way that we have used the word entity<\/em> at various points in the book \u2014 a typical dictionary definition for entity<\/em> is the following:<\/p>\n entity<\/strong>, n. \/\u02c8\u025bn.t\u026a.ti\/ something that exists separately of and can be distinguished from other parts of the world<\/p>\n<\/blockquote>\n (In fact, it is not entirely typical \u2014 most dictionaries use the word thing<\/em> to define entity, which leads us back to the problem just mentioned.)<\/p>\n The problem with this definition is that it is so abstract that it can only be understood in terms of concrete examples: we have to know a range of entities \u2014 people, ideas, objects, plants, animals, etc. \u2014 in order for this definition to make any sense at all.<\/p>\n The same problems exist if, instead of using dictionary definitions, we define words in terms of lists of properties, as we have also done above, for example, when stating the meaning of tree<\/em> as \u2018is a plant\u2019, \u2018is perennial\u2019, \u2018is made of wood\u2019, \u2018has a single trunk\u2019, \u2018has branches\u2019, \u2018has leaves\u2019. Such a list format forces us to think more precisely about the properties that define the extension of a word, and so it can be more useful than a dictionary definition written as a continuous text, but the two are essentially just different representations of the same information \u2014 we can easily paraphrase a dictionary definition as a list of properties or vice versa.<\/p>\n There are two ways of dealing with this problem. The first one would be to simply accept it. When we are describing the meaning of a word, we are using language to describe commonalities of the mental experiences that humans have when interacting \u2014 or thinking about interacting \u2014 with one or more of its referents and the relations between referents. It would be surprising if language would allow us to do so in a way that is free of circularities and free of complex relationships between more abstract and more concrete meanings. Since as linguists we are describing the meaning of linguistic expressions for humans, not for aliens or robots, we can rely on the assumption that those humans will somehow make sense of these descriptions. The circularity and the abstractions are often unproblematic when discussing the meaning of a morpheme or the difference in meaning between two or more morphemes.<\/p>\n The second way of dealing with this problem would be to come up with a restricted set of expressions that cannot and do not need to be further defined but that can be used to define all other linguistic signs \u2014 atoms of meaning, so to speak. This has been attempted repeatedly, usually not very systematically.<\/p>\n One such attempt that you will occasionally see is to postulate semantic features<\/strong> (or semantic components<\/strong>) and define the meaning of words in terms of the presence or absence of these features. Such features should (a) be able to distinguish the meaning of the word in question from other words and (b) be useful for defining many different words. Such features would be organized hierarchically yielding a taxonomy of all words of a language.<\/p>\n You will not be surprised that nobody has ever managed to create such a feature hierarchy for a substantial portion of the vocabulary of a language. To illustrate the complexity of the task, consider the hierarchy we would need to define a small portion of words from the letter B of a dictionary: bear<\/em>, beauty<\/em>, birch<\/em>, boat<\/em>, boulder<\/em>, boy<\/em>, bramble<\/em>. All of them are entities in the sense of the term discussed above. A major distinction, and one that is likely to be useful for many other words, is that between entities that are (potentially) ALIVE (bear<\/em>, birch<\/em>, boy<\/em>, bramble<\/em>) and those that are not (beauty<\/em>, boat<\/em>, boulder<\/em>). Within the first category, there is another major distinction between plants and animals (including humans). We might distinguish them by the feature MOBILE. Within the category of ALIVE\/MOBILE entities, we have to distinguish between HUMAN and non-human animals (although humans are animals from a scientific point of view, we do not feel as though they are and we do not treat them linguistically as though they are). Within the domain of entities that are not potentially alive, we might want to distinguish between those that are CONCRETE (can be seen and touched, like boulder<\/em> and boat<\/em>) and those that are not (like beauty<\/em>), and within the the former, we might want to distinguish between NATURAL and non-natural entities. This gives us the taxonomy of features shown in Figure 6.4.1.<\/p>\n Figure 6.4.1. A partial taxonomy of semantic features<\/p><\/div>\n This taxonomy allows us to distinguish the meaning of the very general expressions human<\/em>, animal<\/em>, plant<\/em>, natural object<\/em>, artificial object (artefact)<\/em>, and abstract entity<\/em>. Not all of these would be used in everyday language, some are still too general.<\/p>\n The word boy<\/em> would have the features [+ALIVE, +MOBILE, +HUMAN], but it would share these features with all other words for human beings. To distinguish it from girl<\/em>, we would have to add the feature [+MALE] or [\u2013FEMALE], depending on which of these we consider basic. To distinguish it from man<\/em>, we would have to add the feature [\u2013ADULT]. Thus, boy<\/em> would be defined as [+ALIVE, +MOBILE, +HUMAN, \u2013FEMALE, \u2013ADULT]<\/p>\n The words birch<\/em> and bramble<\/em> would have to be distinguished from each other (one is a tree, one is a shrub), and from other plants. The features we would need to do so no longer have the generality of the features we have posited so far \u2014 they both have trunks, as opposed to other plants, but [\u00b1 TRUNK] does not seem very plausible as a semantic atom, and we would not need it to distinguish any words other than tree<\/em> and shrub<\/em> (and their synonyms). The fact that trees have a single trunk and shrubs do not poses another problem: we cannot easily specify it in terms of presence or absence. It would be better to have some form of quantification \u2014 [0 TRUNK] (for other plants), [1 TRUNK] (for tree<\/em>), [\\<1 TRUNK] (for shrub<\/em>).<\/p>\n Problems like this led most linguists to abandon the idea that the meaning of words can be described in terms of semantically simple and general features. However, many linguists believe that there are some features of this kind that are useful when modeling meaning, because they pick out aspects of a word\u2019s intension that are relevant to morphological and syntactic rules. We will discuss this idea briefly in Chapter 7.<\/p>\n <\/p>\n\n
\n
\n
\n
\n
![\"A](\"https:\/\/linguistica.info\/b\/lei\/wp-content\/uploads\/2024\/12\/download.png\")