18 replies to this topic

[Harvey]

it appears that in the syntax module, the enclitic prepositions ב־, כ־ and ל־ are not separated from the definite article, when attached to a determined noun, making it impossible to distinguish in a syntactical search between, for example, ביום השביעי with and (hypothetically) without a definite article before יום. (At least, that is how I read the tree in Gen 2:2, for example.) Am I correct?

[Robert Holmstedt]

That is a feature of the actual biblical text first, and thus, necessarily a feature of the syntax (which builds upon the morph text, which builds on the digital biblical text).

It is, frankly, also an ambiguous feature of the biblical text since the vocalization is later than the consonantal text. As James Barr pointed out many years ago, there are numerous cases of the Masoretic prep+article pointing (esp. in poetry) that are probably not original.

But, back to the searching -- since you can mix-and-match the morph-tagging features (e.g., presence/absence of the article] and the syntax features, you can certainly construct a search that distinguishes between one phrase with the article and the corresponding phrase without the article. My apologies for not illustrating it right now, I am simply too busy with other work. Perhaps someone else will jump in a provide screen shots.

Edited by Robert Holmstedt, 12 July 2012 - 03:22 PM.

[Harvey]

Thanks. (I am aware of the likely anachronistic appearance of determined enclitic prepositions in early poetry.)

I've tried unsuccessfully to figure out ways to exclude definite enclitic prepositions. I tried inserting NOT + SPECIFIER within 1 preceding the noun, and I've tried inserting NOT + LEX ׁ(ה); neither seems to work, presumably because the definite article isn't coded as a morph in the word strings in question. I look forward to learning how it can be done.

[Peter Bekins]

Thanks. (I am aware of the likely anachronistic appearance of determined enclitic prepositions in early poetry.)

I've tried unsuccessfully to figure out ways to exclude definite enclitic prepositions. I tried inserting NOT + SPECIFIER within 1 preceding the noun, and I've tried inserting NOT + LEX ׁ(ה); neither seems to work, presumably because the definite article isn't coded as a morph in the word strings in question. I look forward to learning how it can be done.

As Robert said, this is an issue with the double nature of the MT as base consonantal text + vowel overlay. The morphological tagging necessarily follows the consonantal text so it does not recognize a lexical ה and the syntax consequently does not recognize a specifier here. Rather, the morph DB tags these as "compound preposition article" which you will find under the "particle" menu.

Hope that helps,
Pete

[Harvey]

Thanks, Pete. That did it.

I'm working my way up the learning curve, but it's a slow process!

Harvey

[Harvey]

Here's what I assume is a really basic question: How do I search for a noun that is immediately preceded by neither the definite article nor a "compound preposition article"? More generally, how do a I search for an element that is immediately preceded by neither of two other elements?

[Helen Brown]

In general, you create a column before the Noun, drag to it the articles you want to avoid, drag the NOT item over the elements, and specifiy the Within to be 1. Don't check the Search in both directions, and that should do the job.

[Harvey]

 Screenshot.tiff   296.36K   11 downloadsI have been trying to do just that. As you can see from the attached screenshot, I continue to get nouns preceded by the definite article. (I know that this is a silly search, but in the course of a more complex one, I isolated my difficulty to this issue. I also realize that including both NOT-Specifier and NOT-LEX=ה is overkill, but I didn't want to overlook any possible solution.) I look forward to learning where I went astray.

[Peter Bekins]

 Screenshot.tiff   296.36K   11 downloadsI have been trying to do just that. As you can see from the attached screenshot, I continue to get nouns preceded by the definite article. (I know that this is a silly search, but in the course of a more complex one, I isolated my difficulty to this issue. I also realize that including both NOT-Specifier and NOT-LEX=ה is overkill, but I didn't want to overlook any possible solution.) I look forward to learning where I went astray.

Instead of using either lexical ה or specifier, use the morph tag particle > article. I don't think searching on ה is limited to definite articles since there is also an interrogative ה. I am not sure why "specifier" doesn't work, but I am not exactly clear on how the syntax and morph search algorithms interact with each other.

[Harvey]

Thanks, Pete. That did it. The lexeme ha- that I selected was the one glossed "the," but I am learning that what should logically work in searches doesn't always do so. I appreciate the work-around suggestions of more experienced hands, like you.

[Robert Holmstedt]

Whether or not the search results from Accordance are logical depends a good bit on the logic of the search. Your search included some potential illogicality, since how the NOT is applied to multiple items gets rather complex. (Try writing it out in logical form and you'd see the ambiguity.)

The simpler way to get the search you want is to use the shift-click function to specify in the Particle menu that you want article and compound-preposition-article together and then use the NOT function to negate the choice.

 Screen Shot 2012-07-15 at 1.39.01 PM.png   14.34K   10 downloads

Now, if you really need to add distinct items (i.e., those from different boxes, which can't be joined by shift-click in a single box), then you could search for the items (negated, if needed) separately and use the method of searching within the HITS of a previous search, which is described in the Accordance Help. That is, search #1 and then search #2 within the HITS of search #1.

Edited by Robert Holmstedt, 15 July 2012 - 02:49 PM.

[Harvey]

I haven't been able to locate an explanation of searching with HITS in Accordance Help. (It may well be there and I've missed it.) If you could point me to it, I'd be grateful. I assume this is different than Helen's suggestion ("Currently you cannot directly search the contents of a tab, but you can open another tab and search it with [CONTENTS first tab] <AND> my new query") for searching within search results.

Edited by Harvey, 15 July 2012 - 04:47 PM.

[Michael J. Bolesta]

Search the help page: http://accordancefil...OSX/Default.htm

[Harvey]

I did, of course.

[Joel Brown]

The [CONTENTS] command runs a search, requiring that any hits are also located only in the verses found in the linked tab. So, if I have one tab do a search for 'praise' and another tab do a search for [CONTENTS Tab1] <AND> land, the only hits are where land is found, and the verse was found in the 'praise' search. Obviously, this example is unnecessary, but the idea applies.

The [HITS] command works very similarly, but instead of requiring the verses to match, it represents the hit words of the linked tab. So, for example, I could run a search for '*' in Genesis 1 to get all of the words in Genesis 1. I could then in my second tab run a search like [COUNT 1-10]@[HITS Tab1] in Isaiah, to find all words in Isaiah that also exist in Genesis 1, and only occur 1-10 times.

Hope this helps!

[Harvey]

Thanks for this information, Joel. Is there any written guide to available operators and the syntax of searches via the search box? (I've mainly been using the construct window, which is (relatively) self-explanatory.)

Edited by Harvey, 16 July 2012 - 07:02 PM.

[Helen Brown]

The Help files list and explain all the commands. The Getting Started with Syntax has some explanation of the new commands.

[Harvey]

Still struggling here. I am now trying to search for certain occurrences of adjectives while excluding those where the adjective is immediately followed by a preposition. I have tried NOT + particle=preposition in the column following the adjective, WITHIN 1 of the adjective column, and I have tried particle=preposition with NOT checked in the particle drop-down menu; but neither seems to do the trick. Is there a different approach that would work?

[Ken Simpson]

Hi Harvey, to do a simple search that finds adjacent terms, just put them next to each other. So if I understand your search:

just try [PARTICLE -preposition][ADJECTIVE]

I did this by putting Accordance into "Words" mode, then used the menu item Tag/Adjective then the menu item Tag/Particle/Preposition (ticked not)

That gives 5012 hits in 3959 verses, and while I can't claim to have looked at all the hits, they look OK to me.

Does that make sense?

Ken