1. Why replace the Cohler translation ?
Any translator owes at least something to anyone who has previously taken on the same work. While Thomas Nugent’s freewheeling translation (often in fact a paraphrase) is altogether unacceptable by today’s standards, it is still sometimes useful if only because it is contemporary with Montesquieu and has a good chance of reflecting period vocabulary with reasonable accuracy (at least insofar as Nugent understands Montesquieu’s meaning). This makes it, like period dictionaries such as the Dictionnaire de Trévoux, an important resource because the meaning of words has often shifted in the interim ; in the case of technical (especially legal) terminology, a meaning may have disappeared altogether in 270 years, and sometimes the word as well.
The Cambridge translation of 1989  also has many qualities, including a careful and controlled use of terminology that contrasts with Nugent more nonchalant approach. It solves, or at least tries to solve, not a few knotty problems which Nugent had adroitly dodged, and comes up with a number of quite elegant formulations, some of which are worth imitating.
It nevertheless harbors many mistakes, often bound up with systematic vocabulary choices. Some are just awkward, such as “soften” for adoucir (e.g., XII.30), some are more than slightly off, and some are quite wrong, for example “disgrace” for disgrâce (e.g., XII.30) and “empire” for empire (passim). Many of these problems arise from misplaced confidence in cognates, which, goodness knows, is a familiar devil in translation. Argent does sometimes mean silver, as the Cambridge translation usually has it, but by no means always. Empire does not mean “empire” except when it is a formal designation for a type of political structure, and certainly not in Quand la beauté demande l’empire, la raison le fait refuser (XVI.2). Droit does not always mean “right”, it even sometimes means “law”. Univers does not mean “universe” but “the entire world ; all created creatures, the totality of all beings”  (Trévoux). And so forth.
Though any attempt at an exhaustive catalogue would be fastidious and needless, I will give just a few examples to illustrate misconstrual of particular terms in specific contexts.
C’est la vexation qui se surmonte elle-même, et se voit contrainte à une certaine douceur (XIII.11). > “This harassement defeats itself and one sees it constrained to be somewhat gentle.”
Se surmonter often has a positive meaning, here something like “surpasses itself”. The resulting sentence is all but incomprehensible (frankly, on occasion, the fate of any translation).
Les hommes s’y soumettent à des maux incroyables (XIV.3). > “Men there suffer unbelievable evils”.
Mal (plural maux) has many uses, and here, in the context of fakirs in India, means pain, not evil.
Les fibres qui n’ont qu’une action très faible et peu de ressort, ne s’usent guère (XIV.10). > “The fibers, which have only a very weak action and little spring, are scarcely used […]”.
To translate user as “use” is a serious oversight. This is a mechanical comparison, where s’user means to wear down or wear out.
[…] il faut que l’esclavage soit pour l’utilité, et non pas pour la volupté (XV.ll). > “[…] slavery must be for utility and not for voluptuousness” [also passim].
Volupté is an annoying word for translators because (like pudeur) it has no very accurate English equivalent. But though it always related to sensuousness, “voluptuousness” is a desperate choice and will not do.
Il en fallut même de terribles pour établir la sûreté de ces maîtres cruels (XV.16). > “There had even to be terrible laws in order to establish security for these cruel masters”.
This passage is not about “terrible” laws but about ferocious laws. Terrible/“terrible” is a tempting but often misleading cognate.
[…] on vit sans cesse le frère, l’oncle, le neveu, que dis-je, le fils, le père, conspirer contre toute sa famille (XVIII.29). > “[…] the brother, the uncle, the nephew (what can I say ?), the son, the father, were seen constantly conspiring against the rest of the family.”
Hard to handle stylistically, but it must first be understood that que dis-je ? does not mean “what can I say ?” or anything close to that ; it raises the rhetorical level by signifying an intensification in the level of allusion.
[…] tout citoyen y aurait sa volonté propre, et ferait valoir à son gré son indépendance (XIX.27). > “[…] each citizen would have his own will and would value his independence according to his taste”.
Faire valoir simply does not mean “to value” ; here, as in too many other passages, the translators were dealing with an idiom they clearly did not understand. Here it means to stand up for, to argue, to assert. But as I said, this sort of accident also befalls Nugent, and without any doubt the present translator as well, at least sometimes.
2. Approach to some general problems
Certain endemic grammatical traits of French are all too familiar to every translator. Among them is the very fluid and multiple use of the pronouns/prepositions/adverbs y and en. Any attempt to retain each and every occurrence in a translation betrays an inexperienced (or machine) translator. Less often noticed is on, an all-purpose pronoun which is exceedingly prevalent in L’Esprit des lois, with close to 7000 occurrences. We may as well concede that this is a text constructed to a substantial extent on the very liberal and flexible use of the pronoun on.
“One” is, of course, an approximate equivalent in English ; the problem is that it cannot sustain prolonged use without appearing stilted, and when overdone it becomes unbearable. A whole repertory of work-arounds is needed which to a word-for-word purist will seem a deviation from the original. But really they are necessary if the text is to sound anything like natural in its target language.
In practice, what translators do, of necessity, is to find other ways – and certainly more than one is needed – to avoid excessive repetition of “one”. Inasmuch as on often in fact means “we”, the first-person plural is then a more than acceptable substitute. Sometimes we can say “a person” or “someone” to translate on, or even “people” when, as on often does, it seems to be a broad generalization. Another useful device, but one not suited to all situations, is the passive voice : instead of on dit que…, we can opt for “it is said that…”
Reasonable consistency of terminology is of course important in any text as complex as L’Esprit des lois – especially, perhaps, an historical or philosophical text – but it is not the only consideration. Montesquieu has many passages constructed on rather strict parallels, very often further reinforced by parallel syntax : in such instances, it is obviously important to adhere to that intention – even when it seems quite rigid – insofar as possible. On the other hand, attempts to apply one English word, and only one, throughout for a given French word inevitably force the translator to distort, or frankly mistake, the meaning in particular instances. I have striven for a reasonable compromise which will unquestionably displease some, particularly when they are thoroughly accustomed to one of the translations already available. But the idea that one can preserve the purity of Montesquieu’s meaning, or anyone else’s, by sticking to the same English equivalent at all times seems to me linguistically misguided, and can induce what might be called category mistakes. Does Montesquieu mean the “same” thing every time he uses a multifunction word (like, say, mœurs) ? Of course not. It is an illusion to thing that any word in one language can occupy the same identical semantic space as any given word in another.
The most obvious change I am incorporating in my translation is in the title, where for a number of reasons it seems to me that the singular Spirit of Law may be a better reflection of Montesquieu’s overall meaning than its traditional rendition. For an initial discussion of this and some of the other problems identified, I refer the reader to my article “On the Nugent translation of L’Esprit des lois” in History of Political Thought, vol. XXXIX, no. 1 (2018), p. 83–106.
I am also breaking in other ways with past translators, even with respect to some terms that appear central or any cas appear frequently in Montesquieu’s arguments. One of these is to substitute something for the “spring” (for ressort) of the various types of governments : this, to me, is a good example of a category mistake. The major objection to “spring” is simply that, while Montesquieu does in a few instances use ressort in a mechanical sense (especially in the 1757 Avertissement), the English word “spring” unlike ressort has no recognized metaphorical value that can “naturalize” such a supposedly literal translation  The use of ressort by Montesquieu is in no way a neologism, but it is in English when it comes out as “spring”. After experimenting with various substitutions I have adopted various solutions related to notions of impetus, incentive, and energy.
Peuple is an exceedingly complicated and ambiguous term in French because it can mean either the body of the people (populus in Latin), in which case it can also take the plural ; or to commoners (plebs) as opposed to upper classes ; or again it can, as “people” often does in English, refer to a polity or aggregate of citizens. “People” is ambiguous also in English, but not in analogous ways. Populaire, as in gouvernement populaire, is closely linked to this latter meaning. To use “people” many times in succession in the singular begins to sound odd. Like previous translators, I have usually in such cases shifted into the plural, despite some slippage of meaning that seems to me sometimes unavoidable. The idea of plebs, also much used by Montesquieu, refers to those who are not noble or rich, and is often best rendered by “commoners” or “common people”. Indeed some other words such as gems and on also are on occasion best translated as “people”.
Pouvoir and puissance have a similar kind of overlapping usage. Sometimes, but by no means always, Montesquieu keeps them separate, even in opposition, and at other times he seems to substitute them indiscriminately (XI.6). I have generally rendered pouvoir as “power” and puissance as “authority”. There are also, of course, occurrences of autorité that also seem to call for “authority” – but not always.
3. A particular note on mœurs
Everyone knows what a problem mœurs is to translate into English. It is indeed tempting to use mores invariably, as the Choler group did in the Cambridge translation of 1989. But it will not do, and a few examples from Montesquieu’s text will show that its meaning in context is not always the same. Consequentially, the strictly consistent use of any term ends up skewing the most obvious sense of some passages. There are 241 occurrences of mœurs in L’Esprit des lois, so the is a matter of some importance.
For one thing, mœurs often occurs in the same immediate context as manières and/or coutumes, so we have to assume at least an element of contrast, even between synonyms. The same is true for such words as lois, of course, of morale (XVI.14) and of police(XXI.1). But what does it mean for mœurs to be pures, or corrompues, combinations found numerous times in this text ? There are several mentions of la bonté des mœurs ; and what exactly can be meant by manquer de mœurs (VI.9), or for a citizen to perdre ses mœurs (XIX.16) ? There are some contexts where mœurs strongly seems to denote conventional morality, as when mœurs are assimilated to continence (VII.8). Pureté de mœurs with respect to women (VII.13) pretty clearly means chastity and nothing else. Yet other contexts are much too political to suggest such an implication, as when Montesquieu speaks of une République fondée sur les mœurs (V.19). Similarly, what could it mean to be gouvernés par les mœurs (VIII.9) : tradition, simply ? An equally difficult question : can donner des mœurs (an expression which appears quite a few times) be defined with any degree of precision ?
The definition of mœurs in the Dictionnaire de Trévoux is carefully written, and very broad :
mœurs, s. f. plur. Façon de vivre, ou d’agir, bonne ou mauvaise ; habitudes naturelles, ou acquises, pour le bien, ou pour le mal, & suivant lesquelles les peuples, ou les particuliers, conduisent les actions de leur vie. Mores ; boni vel mali habitus mentis.
It puts the accent on habits or practices, at first neutral but then quickly suggesting, by the examples given, moral vectors : “La morale traite des mœurs pour les rendre bonnes : elle ne les souffre ni mauvaises, ni indifférentes.” The numerous subsequent examples resemble the contexts in which we find the word in Montesquieu, and there, significantly, the Latin control word (Trévoux being a bilingual dictionary) is no longer mores but mos or usus :
Coutume, usage, façon, manière de vivre grossière ou polie. Mos, usus, agendi ratio. Nous prenons d’ordinaire les mœurs de ceux que nous fréquentons. Port-R. L’amour sans dérèglement adoucit les mœurs les plus sauvages. M. Sc. Les mœurs d’aujourd’hui civilisées et adoucies, ne peuvent souffrir ce qu’elles avaient de farouche autrefois. S. Évr. Il faut avoir des mœurs accommodantes. M. Esp. Platon adoucit et polit les mœurs des Grecs. Dac. Admirez cette diversité de mœurs et de coutumes, qui ne divise pas moins les nations que la différence de leur langage. S. Évr. Socrate faisait profession d’enseigner la science et la doctrine des mœurs. Il instruisait la jeunesse aux bonnes mœurs. Selon nos mœurs veut dire selon nos usages, nos coutumes.
The definition in the Encyclopédie is rather similar :
[A]ctions libres des hommes, naturelles ou acquises, bonnes ou mauvaises, susceptibles de règle et de direction. Leur variété chez les divers peuples du monde dépend du climat, de la religion, des lois, du gouvernement, des besoins, de l’éducation, des manières et des exemples. À mesure que dans chaque nation une de ces causes agit avec plus de force, les autres lui cèdent d’autant.
These definitions led me to think about two possibilities which I had not entertained earlier as English equivalents : “behavior” and “habits.” “Customs” would in principle do as well, but there are too many instances in L’Esprit des lois of coutûmes with a specific, legal meaning. Habitude, however, is a word that does not occur anywhere in the text, so “habits” would not get in the way. But often “behavior” will do even better, as a neutral term suiting a variety of contexts and also staying out of the way of most near-equivalents. There are many passages that string together mœurs, manières, coutumes and other terms that supposedly can be distinguished from each other but really seem nearly synonymous.
Still, there are many occurrences of mœurs in which “behavior” either isn’t a good syntactic fit or sounds strange. I also tried out “standards”, “norms”, and also “ethos”, which sometimes seems to work better than the alternatives, but not in every instance. It is, I know, a nineteenth-century word, still in my view better than mores, which though a Latin word was not re-introduced into English usage until the twentieth century. But another, grammatical problem arises : for “ethos” is necessarily singular while mœurs is necessarily plural, so all the pronouns and verbs have to be swapped out to match. Not that this factor need be absolutely decisive, for it is a mistake to think that the number of the original must always be preserved in the translation.
All this makes it clear enough that the different uses of mœurs support a variety of meanings, not just one fundamental one. And while the dictionaries make it clear that mœurs may be attributed to a person, that is not usually the level on which the notion interests Montesquieu, who generally refers rather to collective mœurs as practiced or manifested by a whole polity. If mœurs are just habits, then that term still must be understood collectively rather than individually. An interesting contrast is made in The Spirit of Law, XIX.14 : “We have said that laws were particular and precise institutions of the legislator, and the ethos and the manners institutions of the nation as a whole.”  This is not the only text that emphasizes the textuality of laws as opposed to the conventionality, the unwritten nature of mœurs. One might even be tempted in this case to translate “memes”, were that term not even more anachronistic than "mores". But mœurs, while not exclusively in the private domain (as opposed to public laws), do relate to the person as opposed to the citizen, and thus are in a sense moral assets or qualities :
Ethos [les mœurs] and manners are practices which the laws have not instituted, or could not, or did not wish to.
One difference between the laws and the ethos is that the laws more determine the acts of the citizen, and the ethos more determines the acts of the man. One difference between the ethos and manners is that the first has more to do with inner conduct, the latter with exterior.
Sometimes, in a state, these things are conflated. Lycurgus made a single code for laws, ethos, and manners, and the legislators of China did likewise. [...]
We must not be surprised if the legislators of Lacedæmon and China conflated laws, ethos and manners : it is because the ethos represents the laws, and manners represent the ethos.[XIX.16] 
This would seem to be a capital text, suggesting again that mœurs means practices or habits, but leaving us with the conundrum of how laws, manners, and mœurs can be understood to “represent” each other.
In all, what works best seems to depend a great deal on the immediate context, including the syntax of a given occurrence.