In his commentary on Revelation, the classic dispensationalist John F. Walvoord gives the following guidelines for how Revelation should be interpreted:
In contrast to the other approaches to the book of Revelation, the futurist position allows a more literal interpretation of the specific prophecies of the book. Though recognizing the frequent symbolism in various prophecies, the events foreshadowed by these symbols and their interpretation are regarded as being fulfilled in a normal way. Hence, the various judgments of God are actually poured out on the earth as contained in the seals, trumpets, and vials.
This type of assessment is typical of dispensationalists, and reflects the wider modernist reaction to the excesses of medieval exegesis, with its “spiritualizing hermeneutic” and love of allegory. According to Walvoord and others, interpreting Revelation’s imagery in a more or less “literal” way is the only way to keep ahistorical spiritualizations at bay and stay “objective” in our approach. Thus Walvoord equates the literal interpretation with the events of Revelation being fulfilled in a “normal” way. In other words, normal must mean literal in his eyes, despite his concession of the “frequent symbolism in various prophecies”. Walvoord is proposing a hermeneutic whereby non-literal language is not acknowledged unless it is virtually inescapable. This becomes clear in the next quote:
In many instances, where symbols are explained in the book of Revelation, they establish a pattern of interpretation which casts a great deal of light upon the meaning of the book as a whole. This introduces a presumption that, where expressions are not explained, they can normally be interpreted according to their natural meaning unless the context clearly indicates otherwise.
Another dispensationalist, Robert L. Thomas, gives a similar assessment in his own commentary on Revelation:
One may wonder how a book of symbols and visions such as Revelation can be interpreted literally. This is not so difficult to understand if one keeps in mind that the symbols and visions were the means of communicating the message to the prophet, but they have a literal meaning unless otherwise indicated in the text. They do not furnish grounds for interpreting the text in a nonliteral fashion. They are to be interpreted as one would interpret the rest of the Bible… So the literal interpretation is the assumption unless something in the text indicates otherwise… The literal approach is fair and consistent.
Each in their own way, Walvoord and Thomas are expressing the common dispensational mantra, “literal if possible”. But is this principle really safe? Is it fair and consistent, as Thomas insists? Does it really bring us consistently closer to John’s intention, or closer even to what his ideal readers would have understood by his visions? What counts, in Walvoord’s mind, as a symbol being “explained” in Revelation? If we are supposed to interpret the text according to its literal meaning unless the context "clearly indicates" otherwise, then what exactly counts as clearly indicating otherwise? And is it even true that the literal meaning is the most “natural” meaning, or is that just the projection of a modernist ideal onto an ancient document which actually reflects a very different worldview?
In fact, there are many symbolic expressions that are never clearly “explained” in Revelation. For instance, the context never clearly indicates that Jesus is not actually a lamb, or that the Holy Spirit is not a menorah, or that the empire and/or antichrist is not actually a beast. But despite our predisposition towards literal interpretation, we all understand these expressions to represent something other than the images used to convey them. Why? Because they are obvious to us.
But given the fact that these few examples already force us to make exceptions to the rule of the literalist method, what makes us so confident that there aren't other instances of unexplained symbols in Revelation? No one expects a seven-headed monster to emerge out of the sea as the fulfillment of Revelation 13, but why do many still expect a horde of demonic locusts as the fulfillment of Revelation 9? If we acknowledge the one to be a visionary symbol, why do we still assume the other to be a direct transcription of history written in advance? Likewise, no one imagines that the “dragon” of Revelation 20 is a literal description of Satan, but then why do we assume that the “pit” is a literal location in which Satan is to be literally imprisoned?
Clearly, the dispensational hermeneutic doesn’t work here, as it is neither fair nor consistent. In tested practice, the principle of “literal if possible” turns out to be an extremely blunt instrument and ironically fulfills its own fears of subjectivity. Hence the shrewd assessment of Vern Poythress:
Literalists understandably fear the introduction of uncontrolled subjectivity, if we are no longer certain what items are nonsymbolical. But in fact it is just as subjective to impose a pedestrian, nonsymbolical reading on a visionary genre to which such reading is alien. Regardless of how far they go in identifying symbolic figures, many interpreters still are captivated by the principle of “literal if possible.” Such a principle may seem safe, and indeed it works well as a first approximation for historical narratives and New Testament letters. But with respect to Revelation and other instances of apocalyptic literature it constantly inhibits interpreters in practice from doing justice to the pervasively visionary character of the discourses. Instead of accepting the visionary, symbolic medium as a natural form, they constantly fight its own inner integrity by requiring explicit proofs of symbolism for each separate, individual vision.
Like many other conservative Christians, I used to view any talk of “symbolism” with great suspicion, supposing that if we allowed the language of Scripture to be interpreted non-literally we would then lose all hope of ever getting at its true historical meaning, because its meaning would be wholly in the hands of the interpreter and we could make it mean virtually whatever we wanted it to. The possibilities would be as endless as our imaginations. Allowing symbolism, I thought, was a slippery slope leading straight down to complete hermeneutical subjectivity. And, to be sure, such fear is justified to some degree by the ahistorical way that many have interpreted the symbolism of prophecy throughout the history of the church.
In hindsight, however, I’ve come to realize that this was an extremely ironic stance to take, for it assumes that the worldview and writing style of the richly storied Jewish world of the first century was exactly the same as that of the post-Enlightenment West, when in many ways it couldn’t be more different. The irony of this was that my fear was actually self-fulfilling, for it kept me from reading biblical prophecy in the way that it asks to be read, grammatically and historically. The Bible does not ask to be read literally in all cases, rather it asks to be read in its own historical and grammatical context, and many times--especially when it comes to apocalyptic literature like Revelation--that translates to a high degree of symbolism.
 Walvoord, The Revelation of Jesus Christ (Chicago: Moody, 1966) 21
 Ibid. 30
 Thomas, Revelation 1-7 (Chicago: Moody, 1992) 35, 36, 37
 Poythress, “Genre and Hermeneutics In Revelation 20:1-6,” JETS 36/1 (1993): 51