In OTTC there has been a notable shift towards conservatism with regard to conjectural emendation within the past few generations. Most scholars now generally avoid proposing emendations and may even reject them outright as a matter of principle. This means that if you want to find possible emendations for a difficult text you are forced to look at the older commentaries, because recent commentaries offer very few, if any. This recent trend in OTTC is certainly not unjustified, even if it is a bit reactionary. In conjunction with my research on the text of the Genesis Flood narrative, I have been reading some older commentaries and noticing a number of patterns. Anyone who wants to look at the older critical commentaries must use them with caution and discernment, recognizing their many abuses and uncritical (ironic...) treatment of the text. A few categories of such abuses I have noticed might be instructive.
1) Many scholars (Ball comes to mind) are in the habit of emending the text of Genesis to conform it to other Ancient Near Eastern Flood accounts. While ANE parallels may provide helpful context, scholars must never forget that the Genesis Flood narrative is a fundamentally unique literary creation. Genesis has its own narrative structure and rhetorical purposes, which are only obfuscated by faulty emendations. The Genesis account should not be arbitrarily forced to conform to other ANE accounts, because despite their similarities, they are literarily independent of one another. Such flawed methodology treats the Genesis text as if it were simply a textual witness to the ANE accounts rather than a text-critical goal in its own right. Thus, e.g., the sending of the raven should not be moved to after the sending of the dove in Genesis simply because that is the order given in an ANE parallel.
2) Many scholars (Gunkel comes to mind) are in the habit of emending the text of Genesis in support of source-critical theories. Reading their commentaries, I am hit with a constant refrain to the effect that "this word/phrase/clause/verse should be omitted as a gloss, because it is in the style of another source." There are two main problems with this. First, these statements betray a misguided goal for their critical texts, also evidenced by the very layouts of their commentaries. These scholars are not interested in establishing the text of Genesis as a completed literary whole, but rather are only interested in Genesis for its supposed testimony to earlier sources. They freely omit within sections assigned to one source any phrases that fit the style of another source, because they do not reflect the "original text" of the source. These emendations are inappropriate for the textual critic of Genesis, however, because there is normally no way to distinguish between secondary glosses conforming the styles of different sources and the work of a final redactor/author and his finished product, the ultimate goal of OTTC. Second, these emendations betray a fundamental weakness of the traditional source-critical approaches. Despite popular misconceptions (and scholarly assertions), the text of the Genesis Flood narrative does not easily lend itself to reliable separation into prior source material. The reality is that "J" sections are so pervasively "glossed" with "P" style and vice versa that the reliability of the source critics' stylistic criteria may be justly criticized as not being obviated by the extant text. Source critics, starting with source divisions of sections determined first and foremost by the usage of the divine names (notably according to the MT only, though that is a whole other discussion), freely omit inconvenient textual evidence that does not fit into their preconceived theories. Textual critics interested in the text of Genesis should be most weary of these biased emendations, forced upon the text as they are to support theories, rather than proposed with due text-critical warrant.
3) Many scholars are in the habit of emending the text of Genesis based on uncritical reconstructions of the ancient versions. Ball, for instance, prefers the text ויצא יצוא ולא שב "And he went out and did not return" in Genesis 8:7 based on the LXX and Syriac readings which could be literally so reconstructed, thinking that it would be most reasonable for the raven not to return to the ark. And yet the chances that the LXX and Syriac actually preserve a true textual variant are slim to none. The chances that this variant is actually original are absolutely nil. It is clear that the LXX and Syriac are exegetical translations of the somewhat ambiguous MT text ויצא יצוא ושוב "And he went out going to and fro(?)" This is especially obvious when one realizes that the Samaritan Pentateuch and a Qumran manuscript understand the MT phrase differently and alter it to mean that the dove did return! The same is true of many difficult words, which the versions seem to struggle to understand and translate. Older scholars tried to reconstruct (and sometimes even adopted into their critical texts) readings based on literalistic, uncritical treatments of the versions that clearly do not reflect true variants. Textual critics should be cautious of many of the preferred emendations in the older commentaries which misuse versional evidence, because in reality they are simply baseless conjectures.
4) Many scholars are in the habit of proposing unnecessary emendations without good reason. Older commentators, operating from a pre-Qumran perspective where the oldest Hebrew witnesses were medieval manuscripts, apparently felt a great deal of freedom with the text. Uncertainty about the history and antiquity of the text and doubts about its reliability led commonly to flippant altering of the text to conform to scholars' preferred exegesis, rather than that exegesis being based on the actual textual evidence. Where there is no apparent textual problem, one should be cautious about proposing alternative texts.
5) Many scholars are in the habit of proposing extreme emendations. One scholar proposed adding a polytheistic reference into the Flood narrative, a truly bizarre emendation. Many others tend to offer long and extremely complex emendations, which do not easily account for the origin of the extant texts, even when much easier explanations are readily available.
6) Many scholars are in the habit of emending texts they do not understand. Many difficult texts have been explained by further discoveries or more nuanced studies, showing that the text is preferable as it stands. It seems like the older critical scholars were often quite overconfident in their abilities and failed to see the limits of their understanding.
These are only a few categories of uncritical treatment of the text in older conjectures. I'm sure there are many others which could be enumerated. Even modern scholars, to be sure, are not immune from making such missteps. I am not convinced that these excesses mean that conjectural emendation should be rejected out of hand today, but it does show that these commentaries should be used with due caution and an understanding of their tendencies and characteristics.
What then is a viable situation for proposing conjectural emendations? Conjectural emendations are appropriate when there are reasonable indications that the extant witnesses do not preserve the archetypal text or that the archetypal text does not preserve the text of the finished literary work, which serves as the textual-critic's end goal. These emendations must be able to explain the origin of all the extant texts, must be contextually suitable, must illuminate the passage, must be philologically sound, and usually should explain an apparent textual difficulty in the extant texts. This is easier said than done, I know...