Allow me to make three points about proof-texts:
1. We should never assume that the use of proof-texts implies a cavalier attitude toward those texts in their contexts (unit of thought, chapter, genre, redemptive-historical location etc.)
2. We should never assume that the use of a proof-text implies a method divorced from responsible exegesis. We should never assume that we have no need to reach for the best commentaries on those texts, that grammatico-historical exegesis is irrelevant, that linguistic study is to be suspended. They should be offered and received in full knowledge and responsible use of all available interpretative tools.
3. We should assume the fundamental unity of Scripture, the "consent of all the parts," because of the superintending work of the Holy Spirit.
When the Westminster Assembly presented the Confession of Faith to Parliament, December 3rd 1646, it was further required of them that the "Assembly should attach their marginal notes, to prove every part from Scripture." Warfield noted that the "proof-texts for both Catechisms occupied the Assembly from Novemeber 30, 1647, to April 12, 1648, and were presented to Parliament April 14, 1648."
A few years back a good friend pointed me to John Frame's short section on proof-texting in The Doctrine of the Knowledge of God. I found it helpful, I hope that you will too:
"Proof-texting" has become almost a term of reproach today, but that was not always the case...A proof-text is simply a Scripture reference that is intended to show the basis for a particular theological assertion.John Frame, The Doctrine of the Knowledge of God, p. 197
The danger in proof-texting is well known: proof-texts are sometimes misused and their contextual meaning distorted in an attempt to use them to support teachings they do not really support.
But it has never been shown that texts are always or necessarily misinterpreted when they are used as proofs for doctrines. And after all has been said, theology really cannot do without proof-texts.
Any theology that seeks accord with Scripture (that is, any theology worthy of the name) has an obligation to show where it gets its scriptural warrant. It may not claim to be based on "general scriptural principles"; it must show where Scripture teaches the doctrine in question.
In some cases, the theologian will display this warrant by presenting his own contextual exegesis of the relevant passages. But often an extended exegetical treatement is unnecessary and would be counterproductive.
The relationship of doctrine to text might be an obvious one once the text is cited (e.g., Gen. 1:1 as proof of the creation of the earth), or it may simply require too much space to go over the exegetical issues in detail.
To forbid proof-texts would be to forbid an obviously useful form of theological shorthand.
Obviously, we should not cite proof-texts unless we have a pretty good idea of what they mean in their context. We do not, however, have an obligation always to cite the context with the text, and far less do we have an obligation always to present an exegetical argument supporting our usage of the text.