Symphonic Theology

Chapter 6: Words and Precision

By Dr. Vern Sheridan Poythress

Available by permission of P&R Publishing.

The reader will be glad to know that the print edition is still in print and available for purchase from P&R Publishing and other retail dealers. The P&R website is:

@ 1987 Vern S. Poythress. Published by P&R Publishing Company 2001. Used with permission. All rights to this material is reserved. This material is for personal use only and may not be published in any form without written permission from P&R. This material is not to be distributed to other web locations for retrieval, published in any other form or in other media either in whole or in part or mirrored at other websites without written permission from P&R Publishing Company, P.O. Box 817, Philipsburg, NJ, 08865.

Title Page
Table of Contents

Our ability to use a number of different flexible perspectives is based partly to the fact that terms such as "ethics" or "adultery" or "covenant" or "prophet" can be stretched. We can use such words in a conventional, prosaic way. But we can also stretch them in an imaginative, almost playful way until they give us a perspective on the whole of the Bible.

This flexibility is in fact closely related to the flexibility that occurs in the meaning of words. A key area in our exegesis and our understanding of the Bible is the area of word meanings and the use of words in the Bible. It is also an area in which we can easily make mistakes. 1

Some people have imagined that words in the Bible all have a special technical precision and give us automatically fixed, rigid categories. These fixed categories are then thought to exclude any kind of flexibility in the use of perspectives. In fact, I believe that the opposite is the case. A close look at the way in which words function in ordinary language and in biblical language complements and reinforces what we have observed regarding the use of perspectives.


Each natural language (English, French, Greek, etc.) includes a collection of vocabulary items, or lexical units (words), currently in use. A good dictionary attempts to list all the words and idioms used in the language at the time. This list we will call the vocabulary stock of the language. (Of course, the dictionary may also include obsolete and archaic items representing earlier stages of the language, and there will always be some hard decisions about whether obscure technical terms or ephemeral slang should be included.) What bearing does the vocabulary stock have on interpretation--either interpretation of a biblical text or interpretation of the function of a philosophical category?

This question is difficult. Formerly, individual interpreters had to use their intuition to figure out what was going on in the use of a given word or idiom. With the development of twentieth-century linguistics, semantics (the study of meaning) and lexicology (the study of dictionaries) have taken on a technical character. Of course, there is danger that technicians in their desire for rigor may oversimplify aspects of language. But at its best, the research in semantics provides tools that enable us to sharpen our intuitions, to correct them, and to criticize certain incorrect inferences based on language properties. 2

The main insight to be gleaned for our immediate purposes is an insight into the finite precision of the meaning of words. The vocabulary stock of a language contains only a finite number of words. The speakers of the language are capable of saying an indefinite number of things about an indefinite number of subjects, using these words. This ability is possible because the words themselves, as members of the lexical system, can be applied to a range of cases. For example, the word "pig" applies not to one individual thing alone but to any of a whole class. It applies to porcine animals male and female, young and old, pink and spotted. If we call some creature a pig, we do not say everything we could about the creature but only that it has a collection of characteristics shared by other creatures in the class. We can gain precision by employing a narrow technical term or by inventing one on the spot. A grice, in some dialects, is a young pig. We gain in precision by adding the feature "young." But we simultaneously lose in scope: the word "grice" can be applied in fewer contexts and can appropriately describe fewer creatures. In the technical terminology of linguistics, we gain in "intension" (the number and richness of the features associated with the term). Simultaneously we lose in "extension" (the number of entities to which the term applies).

Moreover, the meaning of most words is largely determined by the contrast of the one word with other, related words. The word "pig" functions to distinguish a certain group of animals within the class of domestic mammalian animals: it contrasts primarily with "dog," "horse," "cow," "cat." But in another language and culture, if people are intensively involved in raising pigs, there may be several words to describe pigs of various kinds. Even in English, a semi-technical vocabulary used primarily by those working with pigs includes "swine," "boar," "sow," "shoat." Each of these words contrasts with the others, so each can be more precise.

These same considerations apply to our understanding of the use of key categories in philosophy and theology. Consider, for instance, the philosophical categories "cause" and "mind" and the theological categories "miracle," "revelation," and "regeneration." How precise are the meanings of these words? We can expect here that there will be a trade-off between precision and scope, between intension and extension. If we make the word precise, it will cover fewer cases.

For instance, we can use the word "miracle" in a vague, general way, as many people are accustomed to doing in everyday English. Any extraordinary event that helps someone out is a "miracle." If we use the word in that way, it has a wide scope, or extension. Many events are miracles. But it has little precision, or intension. It does not say very much about the events, except that they are extraordinary and that they helped someone. On the other hand, we can establish a technically precise meaning for "miracle": for example, a "direct, extraordinary, visible act of God, conveying special revelation, delivering his people, and confounding his enemies." In this case, the word "miracle" has a much smaller extension. Fewer events count as miracles in this technical sense. However, there is a richer intension. We provide quite a few important characteristics of those things we call miracles. Similarly, "revelation" can cover everything that people come to know, or it can cover only the speech of God in the Bible. In the first use, it has wide extension (scope) and little intension (precision). In the second use it has narrower extension and richer intension. "Regeneration" can be used loosely to cover the entire experience of conversion, or more narrowly to describe the initial act of the Holy Spirit in giving a person a new heart.

There are, then, many cases in which we have some choice about how widely or narrowly we are going to use a term. If we define the term widely, it will be vague. We will probably not be saying anything profound. If we use it narrowly, it will cover only a few cases and therefore may not be so useful for what philosophy and theology wish to do in the way of making statements of great generality.


This trade-off is almost always present in ordinary language. But there is one important exception that does crop up in the technical sphere. An exception occurs when there exists a "natural class," or a very numerous group of entities, all or nearly all of which have a large number of common properties. Pigs are a natural class. So are dogs. Electrons form a natural class.

A natural class arises from a state of affairs in the world, not in human language. Because there are natural classes in the world, we can invent technical terms such as canis familiaris or "electron," which are both general and precise. Such technical terms are general in the sense of having wide extension (they apply to many individual entities) and are precise in the sense of having rich intension (anything denoted by them has many specified properties). 3All of the natural sciences use key technical terms. Usually, these terms correspond to natural classes, in which I include not only groups of living things (species, genus, family, etc.) but relational entities such as "force" in physics and "ionic bond" in chemistry.

The key question is whether similar natural classes exist in philosophy and theology. At the very least, these disciplines do use technical terms. Do such terms correspond to natural classes? Or is there a simple trade-off between intension and extension?

There may be some natural classes of a loosely linked sort in philosophy and theology, that is, certain clusters of properties that tend to occur together. 4 But most of these connections are loosely drawn. Our work with multiple perspectives easily enables us see that many times there are multiple connections in a variety of directions. God is in some ways like a father, in some ways like a king, in some circumstances like a warrior, and so on. No one of these analogies is the only correct way to describe God. There is no one natural class (say, the group of all fathers) that uniquely captures who God is.

Philosophers and theologians have not always taken account of the looseness of their categories. Let us take a specific example from Immanuel Kant. Kant argues for the impossibility of substitutionary atonement in the following way:

This debt [of radically evil disposition] can never be discharged by another person, so far as we can judge according to the justice of our human reason. For this is no transmissible liability which can be made over to another like a financial indebtedness (where it is all one to the creditor whether the debtor himself pays the debt or whether some one else pays it for him); rather is it the most personal of all debts, namely a debt of sins, which only the culprit can bear and which no innocent person can assume even though he be magnanimous enough to wish to take it upon himself for the sake of another. 5

Kant's argument depends on the presupposition that certain kinds of debt belong together in natural classes. According to Kant, financial debts are one natural class. They have uniform properties and can always be discharged in certain ways. "Personal" debts are another natural class, to which certain other uniform rules apply. In addition, Kant assumes that possible acts of substitution by an innocent person also belong to a natural class, so that it is possible fairly quickly to draw universal conclusions. In every case, according to Kant, substitution is illicit.

But Kant's arguaments are no better than the analogies that he sets up. Is the "debt" of radical evil really analogous to personal debts and not financial ones? Is it analogous in just the way that Kant suggests? We can answer these questions only if there are a whole cluster of properties of personal debts that always go together. Moreover, is substitution always impossible? It appears from biblical revelation that both Adam and Christ have a unique representative relation to a whole race of people (Rom 5:12-21). Interchange is possible here in a way not quite like interchange between two individuals chosen at random. In sum, Kant's argument depends on the fact that key terms such as "personal debt" and "substitution (by an innocent person)" form natural classes, with both broad extension and rich intension. Kant's arguments become problematic once we question Kant's postulated natural classes.

In the realm of philosophy, almost any chapter of a major philosophical work reveals problems of this kind. Philosophers want to arrive at profound conclusions. Such profound conclusions are possible, however, only if the terms at issue have rich intension and if such intension corresponds to a large natural class. How could one establish the existence of such a natural class? Philosophers almost always attempt to do so by some combination of general reasoning and a few key examples. Kant uses the key examples of financial debts, personal debts (moral crimes?), and the possibilities of substitution in human affairs of justice. But we notice that the examples have become perspectives on the whole. In Kant's case, they become perspectives on what God can do in any case whatsoever. The extension into a perspective is not rigorous. And such an extension can be done by using other categories that would lead to different conclusions. For example, in Romans 5:12-21, Paul uses the example of Adam, which helps to illustrate the substitutionary significance of Christ, even if it does not provide the type of rationally transparent explanation that Kant was looking for. In philosophy and also in theology, this situation is typical. My rule of thumb, therefore, is to question profound conclusions in philosophy or theology that depend largely on a key technical term or system of terms.

There are some other pitfalls in a philosophical approach. The progress of science has shown that guesses about natural classes must be checked and revised again and again by empirical means. For example, consider the development of theories about the origin of human sickness. In ancient Greece and Roman physicians thought that sickness was related to imbalance in the four humors, blood, phlegm, yellow bile, and black bile. These four humors and imbalances among them were therefore treated as natural classes. The discovery of the bacterial source of disease showed that this original guess about natural classes was worthless. Bacteria did actually possess the properties of a natural class. But further refinements in the understanding of disease were still necessary. Continued research has made the picture more and more complex. Viruses, for example, have some properties similar to bacteria and other living things and other properties similar to nonliving things. The discovery of viruses therefore made it clear that a simple division into diseases caused by living organisms and diseases caused by nonliving things would be too simple. Moreover, diseases like cancer are sometimes caused by a virus, but other times they are caused by inorganic cancer-causing substances, and still other times by radiation. We could never guess at this complexity without doing empirical research, which shows again and again that our postulated natural classes are insufficient to explain everything.

Guesses about what natural classes exist are important, but ultimately we cannot second-guess God. We must look at what God has actually done, not simply reason until we think we know what he must have done. In the example from Kant, Kant thought he could figure out what God must have done to handle people's sins. He did not submit himself to God's revelation of what God has in fact done.

The remedy, apparently, is for philosophers not simply to reason things out in isolation but to do check carefully the nature of the world. In so doing they will function more as linguists, psychologists, and theologians by turns. There is very little independent ground left for the philosopher as philosopher.

We can put the matter another way. The natural classes familiar to science are kinds that could have been otherwise. God could have created a world in which an animal looked like a pig and barked like a dog. We have to look at the world to find out what he actually did create. How can philosophical or theological reasoning claim to pronounce on what must be, when the very categories used in the reasoning may or may not correspond to natural classes?

Theology, of course, has resources not open to philosophy. Those who believe that, what the Bible says, God says, may claim to know what must be, not because they have exhaustive insight or because they have surveyed the whole universe, but because God knows exhaustively and has told them what is universally true.

While this confidence is right in principle, we need to reckon with the fact that the Bible is written in ordinary language to ordinary people. God is not mainly concerned in the Bible to furnish grist for the mill of theological experts or speculators. He intends mainly to bring us to know him personally, to save us, to enable us to serve him from our hearts. Hence, very few if any individual words occurring in the Bible have technically precise meanings. 6

For example, we will distort exegesis if we equate hagiazo ("consecrate") with the technical term "sanctification." Sometimes the Greek word designates the perfection in holiness that we have because we are united with Christ (Heb 10:14), which is quite distinct from the progressive growth in Christian maturity and obedience designated by the technical term "sanctification." Similarly, we ought not to equate pisteuo ("believe") and pistis ("faith") in Paul with the technically precise term "saving faith." Sometimes the words are used in other ways. Ungodly people "believe the lie" (2 Thess 2:11). The word pistis sometimes designates the content of Christian faith (Gal 1:23), sometimes a special gift of faith (1 Cor 13:2).

If, then, as systematic theologians we wish to propose a technical term, we must do so by generalizing from a large number of analogous patterns and teachings that we see in the Bible. Theologians have already identified these patterns for the technical term "Trinity" and for many other technical terms. This work can be valuable. But two restrictions should be borne in mind. First, we should be aware that there can be no category system of systematic theology that is more ultimate than the Bible itself, in the full richness of its message. The category system is always selective and partial in its analysis of the Bible. Second, if we use a category to group together in our mind a series of texts, we thereby put into the background the differences between the texts and the links that some, but not all, of them may have with an alternative category grouping.

For instance, if we use "covenant" as a technical term to describe all of God's relationships with human beings, we group together all the texts that speak of these relationships. Such a grouping is useful, because we then see some common patterns. But note that our technical term "covenant" applies only to human beings. It does not include God's relationship to the sun, moon, and stars. Hence, when we define "covenant" in this way, we put in the background the relation that the Davidic covenant is said to have with God's "covenant with day and night and the fixed laws of heaven and earth" (Jer 33:25). We are likely to overlook this connection, because it extends to the nonhuman order, outside the boundary that we fixed for the technical term "covenant." In defining "covenant" as we have, we also put into the background the contrast between different kinds of relationships that God established with different people. The Davidic promises hold for ever, which is very different from the failing, temporary character of the covenant of Sinai (Jer 31:31-32; compare Jer 33:20-21). When we use one term for both, we may overlook the differences.

Thus, when systematic theologians define a technical category, they are inevitably selective. In the Bible we confront a complexity of interlocking and multifaceted themes. Defining technical terms and categories cannot reduce this complexity into a pristine simplicity. It will not furnish us with "ultimate" categories.


Another kind of imprecision in language arises because the boundaries of the meaning of a word are not fixed with perfect precision. When is a word "correctly" used? Words can be used in imaginative new ways in creating a metaphor. So it is best to start by asking when we can rightly use a word in a nonmetaphorical sense. Some patches of color are clearly red. Some are clearly not red. But what about a color that is off-red in the direction of orange? How much suggestion of orange must it have before it is not red? Or consider another example. Some living things are clearly animals, and some are clearly plants. But what about the intermediate cases, the protozoa Phytomastigophorea which has with both locomotion (animal-like) and photosynthesis (plantlike)? We could mention other puzzling intermediate cases. The "boundary," beyond which we would no longer be comfortable using a word, is "fuzzy."

We make the same observations in terms of metaphor. When is the use of a word such as "pad" a metaphor? Some uses are clearly not metaphors ("the furniture maker padded the chair"). Others are clearly metaphors ("Bill padded his emotional life to withstand the bad news that he knew would come"). Others represent minimal extensions of meaning or dead metaphors. For example, "he padded his essay with truisms" is a metaphorical extension of the most literal meaning of "pad"--"to stuff." But the use of "pad" in the sense of "extend by adding useless extra material" is so common is that is listed in most dictionaries as a separate meaning. It is no longer a metaphor (though it once was) but a regularly recognized extra meaning. We need to recognize that nearly all (if not all) words have a range of application with a clear center but a fuzzy boundary. 7

The phenomenon of fuzzy boundaries means, for one thing, that categories that separate a group into two parts may often admit of intermediate cases. Sometimes we will be unable to say easily which category the case belongs to. Is a tomato a fruit or a vegetable? In such cases, we must beware of an inappropriate kind of appeal to the law of excluded middle. The law of excluded middle says that, for any proposition A, either A is true or not-A is true. This law, however, may not help to solve complex mixed cases. "Either it is raining or it is not raining," someone says. But it may be misting, something between an ordinary rain and no water at all. 8 "Either he acted righteously or wickedly," someone claims. But a particular human action might be righteous in some respects, wicked in others. "Either Mary forced Sue to do it or she did not." But what if Mary urged, cajoled, threatened, or tricked Sue, or used some other means for trying indirectly to force her? "Either this spot is white or it is not white." What if it is off-white? What if some parts of the spot are more off-white, while others are closer to pure white? What sort of lighting is one to use in looking at the spot? What do we do with this kind of difficulty?

We can eliminate these difficulties only by postulating an infinitely precise boundary for a key term. For instance, we could try to define "white" precisely. We could introduce specific data on the different mathematical frequencies of the light waves. Then, however, we would be dealing not with the English word "white" but with an invented technical meaning attached to the same sound sequence "white." As long as we are using a natural language rather than a formalized language of mathematics, fuzzy boundaries are going to interfere with the ideal of infinite precision.

The existence of such fuzzy boundaries also means that we must be willing to admit that other people can, if they wish, draw the boundary at a different point. It is a semantic question as to whether one wants to call a tomato a fruit or a vegetable or both. Likewise, it is a semantic question whether one wants to apply the word "cause" to God or only to connections immanent in the world. It is a semantic question whether one wants to apply the word "miracle" to extraordinary events in our time. The real problems arise because people often will not recognize any distinctions that differ from their own choice of terminology.

The examples here are closely related to what Ludwig Wittgenstein dubbed "family resemblances." Within an ordinary human family there are resemblances among the various members. But such resemblances are a matter of degree. It may well be that there is no one, obvious property shared by all the members of the family. But even if some such property or properties are shared (say, brown eyes and brown hair), these features themselves may not be a sufficient basis for defining the boundaries of the family. To get a reasonable boundary, we may have to require that each member of the family share these two properties and at least one other family feature.

Similar things may happen in dealing with other types of entities grouped together by using a word as the common label. The entities have resemblances among themselves. They may also have resemblances of various kinds with entities outside the group. What determines what counts as a member of the group? How many different kinds of resemblances must a test case share, and how closely must it resemble the paradigm case in each of these different kinds of resemblance in order for it to be counted as a member of the group? For example, how big must the drops be for mist to be counted as "rain"? There is no obvious, common-sense, agreed answer. People may also differ in their judgments over just when they know confidently what the answer is.

We ought to recognize, then, the fluid character of meaning boundaries and the complex character of resemblances in dealing with language use both outside and inside of the Bible. Such recognition can only be for the good, since it gives us more accurate insight into what the Bible says and how much language tells us about the world.

Philosophy and theology, however, are tempted to retreat from this complexity whenever they are bewitched by the mathematical ideal of achieving the accuracy and certainty of Euclid. But even mathematics, we now know, is not so immune to change or uncertainty as was once thought. 9 Moreover, even the kind of accuracy that mathematics does achieve is achieved by eliminating most of the intension that characteristizes ordinary language. The meaning of a number, a function, or a mathematical structure is almost wholly determined by relationships among the parts or relationships with other mathematical objects in a system. "Three" means little other than a contrast with and ability to combine with "two," "four," etc. in specified ways. As long as key words in philosophy and theology are loaded with human, personal, intentional meaning, the appearance of rigorous precision will be an illusion. Any apparent precision always neglects fuzzy boundaries. Moveover, resemblances that cross over the boundaries often make it possible to imagine drawing boundaries in altogether different ways.


In addition, vagueness in the meaning of words is related to the way in which words are learned. 10 Consider how children acquire the vocabulary of their native language. Almost all words, except the most technical, are learned not by explicit definition but by hearing others use them. Now when we as adults use a word, we do not consciously remember all the instances in which we heard it used in the past. However, we have somehow retained an impression of the overall scope of the word's meaning. We have somehow "averaged" the word's occurrences in the past in order to know how to use it now.

In the learning process, each of us hears a word only a finite number of times. which is not enough to enable us to infer the boundaries of its meaning with infinite precision. We have heard the word "rain" used in clear-cut cases and have had its use rejected in other cases. But that background of experience does not enable us to decide confidently whether the word applies when we encounter an intermediate case, such as a slowly falling mist. Past instances of use of "rain" do not suffice to make up our minds about cases that do not fit the past.

Hence, in general we may expect that appropriate uses of a word will "fade off" into inappropriate uses. Some uses are clearly appropriate; some are clearly inappropriate. But in between there are difficult cases. The boundary between the appropriate and the inappropriate use is fuzzy. Are technical terms, then, an exception? Technical terms may sometimes be defined with the explicit purpose of eliminating fuzziness on the boundaries. But even so, the definition itself will contain words of ordinary English that have fuzzy boundaries. Or it may contain other technical terms that will themselves be defined in terms of words with fuzzy boundaries. Elaborate definitions can, of course, eliminate a great deal of fuzziness in the spots where precision matters most. But they will not eliminate it all.

These observations confirm the point made earlier about limited precision and give us an additional way of testing the cogency of arguments. When philosophy or theology introduces a special word that will play a key role in an argument, we can ask how we are expected to learn the meaning or range of usage of the word and what the boundaries are to its application. Even if we are given a few examples or a few instances of its use, the boundaries may remain quite unclear. We need to ask whether we know the use of the term in a situation very different from what the speaker provides. A new word or an old word used in a new sense remains quite vague unless there is an explicit definition. And if there is one, we can ask where the vague boundaries are in the terms used in the definition itself. Such questions will serve to alert us to fallacies or untested presuppositions that sometimes slip in because of movable boundaries.


  1. See, e.g., D. A. Carson, Exegetical Fallacies (Grand Rapids: Baker, 1984), especially pp. 25-66.
  2. The field of semantics is now a large one, and it is not possible here to consider at length the principles and insights that have been derived. For a fuller discussion, the best starting point is John Lyons's Semantics (2 vols; New York: Cambridge University Press, 1977), which contains full bibliographical resources. For the needs of the average biblical student, this book is still too detailed and technical. Several works have applied the insights of semantics and lexicology to biblical studies, and these books will prove to be the most relevant. The two outstanding ones are James Barr, The Semantics of Biblical Language (London: Oxford University Press, 1961); and Moises Silva, Biblical Words and Their Meaning (Grand Rapids: Zondervan, 1983). Carson's Exegetical Fallacies is an excellent book covering a larger range of topics. The reader should also be aware of the work of Arthur Gibson, Biblical Semantic Logic (New York: St. Martin's, 1981); Anthony Thiselton, The Two Horizons: New Testament Hermeneutics and Philosophical Description with Special Reference to Heidegger, Bultmann, Gadamer, and Wittgenstein (Grand Rapids: Eerdmans, 1980); Ludwig Wittgenstein, Preliminary Studies for the "Philosophical Investigations," Generally Known as the Blue and Brown Books (New York: Harper, 1958); idem, Philosophical Investigations, 3d ed. (New York: Macmillan, 1968). Some of the same points that I make below I have developed in another connection in Poythress, "Adequacy of Language and Accommodation," in Hermeneutics, Inerrancy, and the Bible, ed. Earl D. Radmacher and Robert D. Preus (Grand Rapids: Zondervan, 1984), 349-76.
  3. In making these observations, I am not proposing a revival of nominalism. Nominalism (at least in its naive form) says that words and names are only labels for individuals or collections of individuals, that "red," for example, is merely a label for a collection. By contrast, I would say that God has ordained that some created things have common properties (i.e., some things are red). I am proposing neither nominalism nor realism in the medieval sense.
  4. This chapter aims at generalizations about words. In doing so, it singles out clusters of properties that go together. Though its roots are more in linguistics than in philosophy, it might be considered as a philosophical analysis of certain natural classes (e.g., "properties," "categories," "technical terms," "intension"). However, my conclusions go little beyond confirming the best of what we do by common sense. My terms are not technically precise, otherwise I would be saying a great deal about the clustering of a large number of properties in a striking number of instances. However, without a great deal of precision in the definitions, my conclusions are necessarily more general.
  5. Immanuel Kant, Religion within the Limits of Reason Alone (Chicago and London: Open Court, 1934), 66.
  6. For a full discussion of these problems, see Carson, Exegetical Fallacies, 46.
  7. See Stephen Ullmann, Semantics (Oxford: Blackwell, 1964), 125-127.
  8. I owe this example to Kenneth L. Pike.
  9. See Lakatos, Proofs and Refutations; Kline, Mathematics.
  10. Poythress, "Adequacy of Language."
Subscribe to RPM
RPM subscribers receive an email notification each time a new issue is published. Notifications include the title, author, and description of each article in the issue, as well as links directly to the articles. Like RPM itself, subscriptions are free. Click here to subscribe.