Prefer the Harder Reading is one of the four traditional transcriptional canons, that is, rules about what reading we should prefer based on what scribes supposedly did to the text as they copied it. Prefer the Harder Reading (also known by its Latin term, Lectio Potior Brevior) is based on the reasoning that says that it is more likely that scribes would improve the text or remove difficulties than the reverse. As a result, it advises us that we should prefer the more difficult reading when faced with a case of textual variation.

However, just like some of the other transcriptional canons, this idea seems to fly in the face of common sense. As John Dobson writes, 'This kind of reasoning needs to be questioned. Any error made in copying a sentence is likely to make it more difficult to understand' (Learn New Testament Greek, 3rd ed., p316). The fact is that it is much easier to make a mistake in a manuscript than it is to put it right again; there are hundreds of ways to sabotage complex machines, and certain machines will only work if everything is just right. James Royse wrote: "as one increases one's acquaintance with manuscripts, it becomes clear that scribes make virtually any kind of error imaginable sometime or other" (Scribal Habits, p9). If textual criticism is all about weeding out mistakes from a text, it would seem almost strangely back-to-front for textual critics to be preferring readings that make the text more difficult to understand instead of readings that make more sense.

However, listen to one of today's leading NT textual critics, Bart Ehrman, as he explains the counter-intuitive logic of this canon to the uninitiated in commenting on a textual variant in Mark 1:41. Ehrman argues that we should prefer the variant reading, 'Jesus, being moved with anger, stretched out his hand and touched (the leper)', instead of the reading 'Jesus, being moved with compassion, stretched out his hand and touched (the leper)'.

Ehrman says: 'On the contrary, and this may indeed seems backwards at first, the fact that one of the readings makes such good sense and is easy to understand is precisely what makes some scholars suspect that it is wrong' [1]. This logic sounds almost Orwellian, but Ehrman goes on to say that is that it is hard to see why any scribe would change the word 'compassion' (if it was original) to the word 'anger', while it is easy to see why a scribe might make the change in the other direction. Once the argument is framed this way, Ehrman's logic seems hard to argue with.

How do we tell whose logic is correct, Dobson's or Ehrman's? Easy: we need to ask whether there is any evidence for the idea that scribes tended to improve the text rather than corrupt its meaning. Ivory tower speculation about what scribes might have done is all very nice and well, but is there any evidence-base to justify the apparent logic?


Here are the results from Andrew Wilson's study of scribal habits based on 2279 singular readings in 10 chapters of the New Testament text (Filologia Neotestamentaria, 2011):

  Nonsense Harder Sense Harder Style Neutral Easier Style Easier Sense
Singular Readings 558 244 606 676 188 8
Percentages 24% 11% 27% 30% 8% 0.4%

These results show that while scribes often created nonsense, and readings that made the sense more difficult, they rarely made the text easier. Only 8 readings out of 2279 singular readings (0.4%) showed scribes actually improving the sense of the text or removing a difficulty. This result provides little confirmation for the canon.

Textual critics will find these results hard to believe (accustomed as they are to the traditional preference for harder readings), and easy to ignore (because they are only the result of one study). But the beauty of a scientific (as opposed to speculative) approach to working out what scribes tended to do, is that scientific approaches are based on experimental (that is, verifiable, repeatable, and falsifiable) methods. So, it is possible for anyone to pick out a chapter from the New Testament (or a number of NT manuscripts), identify all the singular readings, and then test the singulars to see whether the scribes made the text more difficult or improved its meaning.

In fact, once we look through the singular readings discussed in Colwell's, Royse's, Head's and Hernandez's studies on singular readings, we find similar levels of scribal improvements.

  Colwell Royse Head Hernandez
Singulars in Study 1014 1125 77 322
Improvements 5 9 0 4
Percentage 0.5% 0.8% 0 1.2%

These results show a similar level of scribal improvement: around 1%. In Royse's study, there are so few harder readings commented upon that the Harder Reading canon is not even mentioned in the subject index, and the term is only used once in the entire text, and not until we get to page 730 in one of the footnotes.

An easy way to test the canon for yourself would be to take Head's two articles,  and go through all of the 77 non-itacistic  singulars, and see whether they make the text more difficult or improve its sense. Anyone can do this.

Clearly, if these figures are true, there is very little reason for NT textual critics to be preferring harder readings.

Now, many textual critics will argue that the harder reading canon is valid and these results are not. However, the burden of proof rests with those who claim that the canon represents reality. Are there any studies which demonstrate the validity of the harder reading canon? The answer is, No. As the current author wrote in his Filologia Neotestamentaria article on scribal habits:

Science has two components: theory and observation (or research). Theory without research is mere speculation; observation without theory is mere stamp collecting

Unless textual critics can produce actual research evidence that demonstrates the validity of the harder reading canon, why should anybody believe in Prefer the Harder Reading. It simply will not do for textual critics to say, Well, every other textual critic believes in it. Such silly claims, however, are basically all that can be mustered in defence of the canon (other than imaginative reconstructions about the way scribes behaved - i.e. speculation).

Once we understand that there is very little evidence for the assertion that scribes tended to improve the text rather than corrupt it, other pieces of evidence make more sense too. Notice that 'harder readings' that are sometimes suggested as possible original readings, like the following, are all singular readings (or nearly so):

  • Mark 1:41 - the 'angry Jesus' heals,
  • Hebrews 2:9 - Jesus dies 'without God' (instead of 'by the grace of God'),
  • John 1:34 - Jesus is the chosen of God (instead of 'the Son of God)


It is noteworthy that such harder readings are found with extreme minority support from the manuscripts. But if scribes were given to improving the text and removing difficulties (according to the traditional canon), surely we would find lots of singular readings (or near-singular readings) that were easier (not harder, as in these cases). Instead, we find lots of cases of harder readings among the singular (or near-singular) readings, just as in the three cases listed above. This again indicates that the traditional canon is wrong: it is far more likely that these readings have been scribally-created. 

Similarly, consider the idea now in vogue, that the quest to ascertain the original text is impossible and misguided. It, too, is usually based on arguments that start with a singular reading in an early papyrus that presents us with a harder reading, or with a harder reading found in a very small minority of manuscripts (see Swanson, for example in his Romans volume, p xxvi, who argues from a singular reading in P25 that the quest to find the original text is not only impossible, but less desirable than the goal of presenting the textual tradition in toto). Swanson calls the 'original pure text' a 'delusion', but sadly, the only thing that deserves to be called delusional is Swanson's mode of arguing from singular, harder readings (which abound in our manuscripts) to his pessimistic conclusion, because studies in singular readings clearly show that scribes regularly introduce novel textual difficulties into the tradition.


Some textual critics will ask how it is possible that scribes could have created such 'harder readings'. Surely, they suggest, scribes would not deliberately create such harder readings. This does indeed seem unlikely, but we often find scribal changes in our manuscripts that are the result of multi-stage changes or corrections, salvage-readings of previous errors, and we may assume that sometimes scribes were even forced to conjecture the correct reading because of the damaged state of their exemplar (particularly in the earliest centuries). In addition, we often find scribes copying on 'auto-pilot' - their minds wandering off the task at hand, as well as their minds wandering off to other parts of the NT. The idea that we may divide scribal changes into two simple categories - deliberate and accidental - is overly-simplistic and not attuned to the realities of what we find actual scribes doing.

Prefer the Harder Reading also conflicts with the most basic rule of textual criticism: prefer the reading that makes the most sense in context. (Modern professional NT textual critics, stuck up their ivory towers, deny this and have substituted another basic principle instead of this rule, however any ordinary 'man on the street' reading a letter from his five-year old daughter with garbled spelling and grammar, or reading a church bulletin with a printing mistake, or reading a troublesome passage in Shakespeare, will put this most basic rule of textual criticism into practice and try to work out what makes the best sense). The basic rule of textual criticism, 'Prefer the Reading that Makes the Most Sense' assumes, of course, that a writer is compos mentis, but the unspoken assumption behind Prefer the Harder Reading is that the writer was either unable to be as clear as he could or was deliberately trying to be difficult. While some writers make it more difficult for readers to get at their meaning on occasion, there is little to be gained from continually trying the readers' patience like this. Any writer who wishes to communicate rather than confuse will minimise difficult readings. This leads us, again, to the same conclusion: we should treat the harder reading as a rare minority.

Further, the preference for the harder reading also frequently clashes with an evangelical view of Scripture. George Salmon, contemporary Regius Professor of Divinity at Trinity College Dublin, commenting upon Westcott and Hort's new textual theory, suggested that 'there was indeed but little rhetorical exaggeration in the statement that the canon of these editors was that Codex Β was infallible and that the Evangelists were not. Nay, it seemed as if Hort regarded it as a note of genuineness if a reading implies error on the part of the sacred writer' [2]. The same is true of many textual critics today: they seem to have far more reverential awe towards ancient scraps of writing material than they have towards the memory of the authors (let alone the Divine author, who is not to be mentioned), for they are content to treat the authors on occasions as the clumsiest, and even error-prone, of communicators. The medium is almost worshipped as infallible (in the case of certain ancient Alexandrian manuscripts), while the message, which has survived even longer on account of the majesty of its subject matter, is frequently allowed to be all muddled-up (by the same manuscripts, which usually have less sensible readings).


Some people will not like the logic of the argument here. Let them find some empirical evidence, then, to defend the Harder Reading canon. New Testament textual criticism has a long history, and no one has yet produced any research evidence for the canon which has been thought worthy of mentioning in any of the handbooks. Whichever way we look at it, then, whether empirically, logically or historically, the Harder Reading canon falls short of the standards we normally expect.

If the traditional canon is wrong, does this mean that we should always prefer the easier reading? No, for there are cases among the singular (and near-singular) readings where scribes will amend the text to produce an easier reading. An example of this is found in Matthew 27:9 where one scribe has changed the word 'Jeremiah' to 'Zechariah' because Matthew's quote from the Old Testament seemed to the scribe more properly to have come from the book of Zechariah. But the fact that only one manuscript contains this easier reading is again proof that most scribes did not try to amend the text to remove difficulties.

The Harder Reading canon is wrong and needs to be retired. Instead, we can say that, as a general rule, we should prefer the reading that makes more sense in context. We should definitely NOT prefer harder readings that are based on limited evidence, just as we should not prefer easier readings if they are only based on very limited manuscript or versional evidence.

We may finish, then, by borrowing the words of one of my friends once he had seen the evidence from singular readings:

Prefer the Dumber Reading always seemed a Dumb Rule.


[1] This comment comes from a lecture: Text and Tradition: The Role of New Testament Manuscripts in Early Christian Studies, The Kenneth W. Clark Lectures, Duke Divinity School, 1997, Lecture One: Text and Interpretation: The Exegetical Significance of the "Original" Text. The lecture transcript can be found at: (emphasis added).

[2] George Salmon, Some Thoughts on the Textual Criticism of the New Testament, (London: John Murray, 1897), p26.




Joomla Templates: from JoomlaShack