The Story of a Misprint: Spinoza’s Ethics 4p66
*SCROLL DOWN FOR UPDATE: I’ve now looked at the Akkerman/Steenbakkers edition and have a few comments on their treatment of this misprint**
I’ve been reading through George Eliot’s translation of Spinoza, now available in a very nice edition edited by Clare Carlisle. The translation is readable, accessible, and beautiful of course, and would be great to recommend to newcomers to Spinoza, except that Eliot didn’t have access to the newer editions of the Ethics. Does this matter?
Carlisle notes some points where the editions from which Eliot worked have been improved upon. But there are many others. Is Eliot’s translation, while an accurate rendition of the editions she had, nevertheless an imperfect reflection of what Spinoza wrote?
Carlisle’s introduction notes that Eliot ‘relied mainly on Carl Bruder’s 1843 Latin of the Latin text, though she also consulted French and German translations from 1841 and 1842, which followed earlier Latin editions’. According to Jelle Kingma:
The examination of nineteenth-century editions of Spinoza’s works leads to the conclusion that most editions are poor in this respect [accuracy]. Paulus and Gfrörer are very inaccurate, Bruder has corrections, but made new mistakes and has doubtful conjectures.
On the other hand, Eliot didn’t use Bruder uncritically. For example, there is a mistake in the demonstration of Part 1, Proposition 4:
The reference to Axiom 4, ‘ax. 4’, makes little sense, since Spinoza invokes it to show that substances are the same as their attributes, and Axiom 4 makes no mention of these. By contrast, Definition 4 explains that an attribute is ‘what the intellect perceives of substance, as constituting its essence’, and seems entirely appropriate here. Eliot corrects ‘ax. 4’ to ‘def. 4’ in her translation.
Interestingly, the man who would become her de facto husband, George Henry Lewes, made the same correction when translating a portion of Spinoza in his 1843 Westminster Review article on Spinoza, and added an explanatory footnote:
The other day I came across a mistake that exists in the original edition of Spinoza’s work — the Opera Posthuma of 1677 — and remains in Bruder. It appears in Proposition 66 of Part 4 of the Ethics. Different translations reflect different ways in which this mistake has been corrected.
The Bruder edition, which Eliot probably used, has:
This contains a clear misprint (‘futura’ for ‘futuri’), which doesn’t exist in the existing manuscripts. But the rest matches the manuscripts, and means:
From the guidance of reason we want a greater future good over a lesser present one, and a lesser present evil that is the cause of some future evil.
This is clearly not what Spinoza wants to say: why should we want a lesser present evil that is the cause of some future evil? Eliot changes the final ‘evil’ to ‘good’, and translates:
Under the guidance of reason we desire a greater good in the future rather than a smaller good in the present, and a smaller evil which is the cause of some future good.
Glazemaker, the Dutch translator, did something similar (‘enig toekomend groter good’) — this was noted by Frederick Pollock in 1880. But this replaces ‘evil’ with ‘greater good’, not just with ‘good’, so it’s not clear that Eliot was using this text. Gförer’s 1830 edition, however, contains exactly the same correction as Eliot made: ‘futuri alicuius mali’ becomes ‘futuri alicuius boni’. Carlisle tells us in her Introduction to Eliot’s translation (p.15) that Eliot looked at Gförer.
But this correction doesn’t help all that much. Now we have Spinoza telling us that we want a lesser present evil, which causes some future good. But Spinoza says nearly exactly the same thing in the Corollary to the proposition. In Eliot’s translation: ‘Reason makes us desire a smaller present evil, which is the cause of a greater future good’. It’s not clear why Spinoza would say the same thing twice. Also he tells us that this Corollary is related to its Proposition in the same way as the previous Corollary is related to its Proposition (65). But that Corollary doesn’t merely repeat any part of its Proposition.
Also, in Eliot’s version of Proposition 66, even with the correction, the second ‘smaller’ is out of place. The first — ‘smaller good in the present’ — is contrasted with ‘greater good in the future’. The second — ‘smaller evil’ — is not contrasted with anything.
In Edwin Curley’s version (drawn from Carl Gebhardt’s 1925 edition), 4p66 runs quite differently:
From the guidance of reason, we want a greater future good in preference to a lesser present one, and a lesser present evil in preference to a greater future one.
But this requires a much more significant change to the Latin. Here is the Gebhardt version, from which Curley worked, compared with Bruder’s (which in this case matches the original 1677 text):
As you can see, ‘quod causa est futur[i] alicuius mali’ has been replaced by a completely different phrase: ‘minus prae majore futuro’.
I wondered where this dramatic change was first made, and on what basis. It appears in the 1883 edition of Jan Pieter Nicolaas Land and Johannes van Vloten:
The footnote there gives two names, Camerer and Schmidt. Land wrote a book in which he criticised earlier editions of the Ethics, and discussed this passage, but social distancing has cut me off from his book — I can only see the bit in the Google Books preview linked. Fortunately a very helpful Twitter scholar pointed me towards the book by Camerer referenced in the footnote (I also had help from David de Bruijn). The book referred to is Theodor Camerer’s Die Lehre Spinoza’s (Spinoza’s Teaching), which has this revealing footnote:
Camerer changes the second half of the proposition to read as it does in the Land and Van Vloten edition. The first reason he gives for this we have already noticed: the correction makes the second half of the proposition analogous to the first. Preferring a lesser present evil to a greater future one matches preferring a greater future good to a lesser present one — they both resist time-discounting.
The second reason relates to our observation that if we make the minor correction to the Proposition (replacing ‘future evil’ with ‘future good’), we end up having the Corollary repeat this point exactly. Camerer’s theory is that a misprint in the Latin was caused by a confusion with this Corollary. If so, it is unlikely that this would have been a transcription error, since it appears in all three witnesses to Spinoza’s (now lost) manuscript. Here it is in the famous, recently-discovered Vatican manuscript:
Perhaps Spinoza himself meant to write something like what Camerer writes for him — ‘et malum praesens minus prae majore futuro’ — and confused himself with what he was about to write in the Corollary. But the fact is that those words weren’t written by Spinoza at all, even if they are constructed on an analogy with what he did write in the first half of Proposition 66. We might choose to suppose that it must have been what Spinoza meant, since what he has written is hard to understand. But it does seem dangerous to completely write over the text Spinoza left us.
In this part of the text, Eliot’s translation is a much better record of the original text than Curley’s.
Not all editorial ‘improvements’ are justified. And Curley might have been tempted towards editions that were too charitable to Spinoza. Perhaps, then, Eliot’s disadvantage, in terms of access to editions, was not really so disadvantageous.
The newest Latin edition of the Ethics, compiled by Piet Steenbakkers and Fokke Akkerman, handles the problem of 4p66 in a slightly different way (you can read about the process on pp.19–20 here). Steenbakkers and Akkerman add two crucial words:
Bonum majus futurum prae minore praesenti, et malum praesens minus <prae bono>, quod causa est futuri alicujus mali, ex rationis ductu appetemus
Kisner and Silverthorne’s translation, based on this edition, renders this:
Under the command of reason we will want a greater future good in preference to a lesser present one, and a lesser present bad thing in preference to a good thing which is the cause of some future bad thing.
So Camerer/Gebhardt has:
- Greater future good > lesser present good;
- Lesser present evil > greater future evil.
- Greater future good > lesser present good;
- We desire a lesser present evil that causes some future good.
And Steenbakkers/Akkerman has:
- Greater future good > lesser present good;
- Lesser present evil > (present?) good that causes some future evil.
In a future post, I will discuss Steenbakkers and Akkerman’s reason for their own choice, and see how Eliot’s alternative compares.
I’ve now had a chance to look at Steenbakkers’ and Akkerman’s edition and have a few comments about their way of dealing with this passage.
Steenbakkers and Pierre-François Moreau claim in their Introduction that the 4p66 misprint is the ‘most painful of the thorny problems’ for Spinoza’s readers (p.65). Their solution, as I said above, is to add the words ‘prae bono’ at a crucial point.
At footnote 255, pp.586–90, Steenbakkers and Moreau explain the editorial choice. It is, they declare, ‘a minimal conjecture, based on the stages of argumentation [étapes de l’argumentation]’. I won’t go through all of these étapes, but it is interesting that the reason for rejecting the Gförer correction (also Eliot’s) is the consequence that ‘the second part of the proposition becomes redundant, since Spinoza has already said it in the Scholium to Proposition 62’ (p.589).
The part of that Scholium they mean is:
Si nos de rerum duratione adaequatum cognitionem habere […] mens […] quod in praesenti bonum esset, sed causa futuri alicuius mali, minime appeteret.
If we could have adequate knowledge of the duration of things […] the mind […] would least of all want what was good in the present but the cause of some future evil [I’m now using my own translations].
That doesn’t seem quite the same to me as what Eliot has in 4p66, that we desire ‘a smaller evil which is the cause of some future good’. There is a difference between saying that we don’t want a good thing that causes future evil and saying that we want an evil thing that causes future good. Not wanting a cigarette isn’t the same as wanting an ab workout.
My worry about Eliot/Gförer’s correction was different: it seems to make that part of 4p66 very close to 4p66c. Replacing ‘mali’ with ‘<boni>’, they become nearly identical:
Malum praesens minus, quod causa est futuri alicuius <boni>, ex rationis ductu appetemus.
Malum praesens minus, quod est causa majoris futuri boni, ex rationis ductu appetemus.
[From the guidance of reason we want a lesser present evil that is the cause of some future good.
From the guidance of reason we want a lesser present evil that is the cause of a greater future good.]
On the other hand, perhaps Steenbakkers and Moreau don’t want to throw stones here, since their own correction makes 4p66 pretty close to a different part of 4p66c:
Malum praesens minus, <prae bono> quod causa est futuri alicuius mali, ex rationis ductu appetemus.
ex rationis ductu […] bonum praesens minus, quod causa est majoris futuri mali, negligemus.
[From the guidance of reason we want a lesser present good rather than a good that is the cause of some future evil.
From the guidance of reason we leave aside a lesser present good that is the cause of a greater future evil.]
In Eliot/Gförer’s version, the relevant part of 4p66c is derived from the relevant part of 4p66 a fortiori: if we want a lesser present evil that causes a future good, then we want one that causes a greater future good.
In Steenbakkers/Akkerman’s version, the relevant part of 4p66c is also only a tiny step from the relevant part of 4p66: if we want a lesser present good rather than a good that causes some future evil, then we will leave aside a lesser present good that causes some greater future evil (and thus have, presumably, a still lesser present good).
Either way there is a fair amount of redundancy in Spinoza’s argument (you can look at Steenbakkers and Moreau’s more detailed arguments about how well they’ve done at preserving the structure of Spinoza’s ‘stages of argumentation’).
Also, neither Eliot/Gförer nor Steenbakkers/Akkerman make the proposition into very elegant Latin. Still we have this stray ‘minus’, which doesn’t compare with anything else in the clause (the ‘futuri mali<boni>’ is only ‘alicuius’, not ‘majoris’).
The Gebhardt version (also Camerer, Schmidt, Land, Van Vloten, etc.) makes the proposition into much better Latin, and also reduces the redundancy. But, Steenbakkers and Moreau point out, it achieves this ‘at the price of a heavy intervention in the Latin text’ (p.590).
But if we don’t want to pay it, it’s worth pointing out that Gförer/Eliot’s intervention is even lighter than Steenbakkers/Akkerman’s. And on the measure of redundancy they seem about level.