Friday, December 12, 1999

The Swedenborg Foundation is contemplating a "Library Edition" of the theological works,

to be translated according to agreed standards and annotated to prevalent scholarly

standards. Jonathan Rose and I have agreed to draft samples for evaluation, mine to be of

Heaven and Hell. The following information was generated by the need for a footnote

explaining Swedenborg's footnotes.

Heaven and Hell, like White Horse and The New Jerusalem and Its Heavenly Doctrine,

contains many references to specific paragraphs in Arcana Coelestia. We can hardly expect

Swedenborg to have had all 10,837 paragraphs of the latter in his memory, which raises the

question of his indices to it--their purpose, their use, and their composition. Anyone who

has looked at the autographs in Volume XIV of the phototype series cannot fail to be

impressed at the labor involved. They were apparently written in two volumes of blank

sheets and cover hundreds of pages, often densely. In the longer version, entries for some

twenty-six of the more common words (such as "charity," "faith," and "the Word") overran

the space he had allotted them and had to be continued at the end of the volume.

Swedenborg carefully noted the page numbers for these continuations, so the trail can be

followed without much difficulty.

What was the purpose of the indices, and how do they relate to each other?

A plausible case for the genesis of the longer index can be made on the basis of

comparison of its references with the cross references in the Arcana itself.

At first, the references in the Arcana itself are to paragraph numbers. ¶ 85 refers to ¶

81, 91 to 87, 117 to 113, and so on. At ¶ 305 (on Genesis 1:23) we have the first break in

this pattern, a reference to verse 19. After a reference to ¶ 245 in ¶ 379, references to

chapter and verse become the rule and references to paragraph numbers the exception. In ¶

594, though, four paragraph numbers are cited as references for bestia ("animal"), and

this list is repeated with one addition in ¶¶ 674, 714, 774, and 803. A single paragraph

reference for avis ("bird") in ¶ 595 recurs in ¶¶ 776 and 803, though there is a chapter

and verse reference for the same word in ¶ 745. It may be worth noting, incidentally, that

at ¶ 207 we have a genuine rarity--a specific reference forward, namely to ¶ 265. This was

presumably inserted when Swedenborg wrote out the fair copy.

The prevalence of chapter and verse references holds true until midway through chapter

eight. From ¶ 893 (Genesis 8:13) on, the strong preference is again for references to

paragraph numbers, and comparison with the first draft index discloses that the great

majority of the references in the text are to be found in the index. Notable exceptions

are arcus in nube in ¶ 1054, nuditas in ¶ 1079, idololatria in ¶ 1363, and coelestis homo

in ¶ 1442.

The indices do more than merely list numbers. They include brief summaries of the content

of the references, and Swedenborg clearly attended to these in his citations. That is, he

did not simply jot down all the numbers available to him, but selected those which were

particularly germane. He also seems to have done some checking, since in ¶¶ 1243, 1262,

1327, and 1447 he corrects the erroneous "1067" of the index (s.v. terra) to the correct


The rationale behind this pattern seems obvious. At first, when the text to which he was

referring was both relatively small and fresh in his mind, it was not difficult to check

back and find the paragraphs to which he wanted to refer. As the work grew, it became

easier to locate these by their association with the very familiar Biblical text. This too

became cumbersome, though, as possible references multiplied, and somewhere about the

middle of chapter eight Swedenborg undertook the laborious task of keeping a running

index. The references continue into the 10,000 numbers, making it evident that once he

started, he followed through.

A check of Swedenborg's annotations to the first thirty paragraphs of Heaven and Hell

yields comparable results. Where the items annotated are individual words, the

overwhelming majority of the references can be found in the index. Sometimes erroneous

references in Heaven and Hell can be corrected. In ¶ 3, the reference to ¶ 5633 (for which

Samuel Worcester in the third Latin edition suggests reading ¶ 5663) clearly reflects the

5638 of the index. When the items referenced are propositions rather than single words,

search via the obvious key words was not productive.

The second index is much shorter and more carefully written. A spot check of some of the

more common terms yielded no reference beyond ¶ 2751 (s.v. Dominus, "Lord"). A careful

comparison of the entries under fides ("faith") with the corresponding entries in the

larger index disclosed that as far as the smaller index went, all but two of its items

could be found in the larger one, and none from the larger was missing. The two

exceptional items were followed by the note vide Persuasio (see "persuasion") and occurred

in the larger index under that heading.

The main difference between the two indices, apart from the obviously greater care with

which the smaller is written, is that there is evidence in the smaller of an effort to

rearrange the entries topically--they are no longer simply in numerical order. Obviously,

there would be no way to do this thoroughly until the Arcana was completed. A kind of

paragraphing and evident change of quill in the smaller index suggests that Swedenborg

wrote it in two phases. On the basis of spot checking, it appears that the references in

the "first paragraphs" are consistently from chapters 1-15--the first volume of the first

Latin edition. The latest reference noted in the "second paragraphs" (¶ 2751, see above)

comes from the interchapter material at the end of Chapter 21--the last chapter of the

second volume of the first Latin edition. It should be noted that in the entries for

Verbum ("Word") and Conjugium ("marriage") the "second paragraphs" do include references

to the first volume--¶¶ 683, 793, and 801 in the former instance and ¶¶ 1719 and 1720 in

the latter.

In the longer index, there is a consistent but less obvious "paragraphing" which also

seems to reflect the end of the first volume, but there is no further such distinction,

even in the entries that run to multiple pages.

We may then reconstruct the genesis and use of the indices as follows. Swedenborg adopted

the discipline of indexing only when the work reached a size that made it necessary for

consistent cross-referencing. Once the first volume was completed, he went through the

index and made a fair copy with the entries sorted as to topic, and he repeated this

process only for the second volume. He continued the strictly sequential index through the

completion of the Arcana, however, and this was therefore available to him for his

annotation of Heaven and Hell.

One puzzle remains. Virtually all of the entries in the longer index are crossed out,

which presumably means that in some fashion or other they have been "attended to."

Sometimes this is done item by item, sometimes with an unbroken line through more than one

item, and sometimes, especially with the shorter entries, by broad diagonal strokes. Two

possible explanations seem equally staggering. The less likely is that he crossed items

off as he cited them in subsequent works, which would mean that somehow or other he

managed to use virtually all of the references he had collected. The other is that

ultimately he completed a fair copy of the entire index, a copy now lost.

contact phil at for any problems or comments