Finding the root of a Hebrew word

 Finding the root is perhaps the most complex task that a new reader of Hebrew discovers. It is a vital task for the beginner to be able to use a lexicon or glossary because all words are listed by root not by word, and the words have prefixes so if you are looking for cawr, or vlawr, or ubawr you won't find them under c or v or u. In fact you won't find u at all since it is equal to v. And of course you won't find Latin letters in any Hebrew concordance except this one. (And this one includes the Hebrew letters as well and they are in the same sequence as the Latin letters! - it was tricky to get this right.)

So job one - to find the root - has its first stumbling block. The second one is that some root letters are weak - and they may disappear or change into some other letter. This is worse than finding the starting letter but there are some shortcuts.

Some letters are reliable and never change (or very rarely). These are: a, b, g, d, z, k, T, c, l, s, y, p, x, q, r, w. (16)

The weak letters are h, v, i, m, n, t. (6)

So you have to remove the prefixes and restore the weak letters, then you can begin to see the root. Finally recognize the suffixes and remove them, but also restoring the letters that may have morphed to allow the suffix to be easily spoken.

These are letters that are used as prefixes: b, h, v, c, l, m, w (rarely in BH), a, i, n, t (for verb forms). Total 11/22.

I developed a morphology routine that works about 99% of the time. Go back to the examples above - 

cawr, or vlawr, or ubawr become c/awr, vl/awr, and ub/awr.

- you can see the root a-w-r - {roughly asher} three strong letters. So the prefixes are c, v-l, and u-b respectively. That one was easy and there were no suffixes. Now go and study the word awr and its uses. Aren't you happy you did that? Notice how rich this root is: a verb, (and its noun), a name of a location or person, and even a very common stand-alone pronoun.

Here's another example: and this shows how to use the vowel column:

vadmtnu    _:_a-_:_a_ei_u becomes v/adm\tnu - not quite perfect if we were thinking that the root was admh - but close enough to find it.

The prefix is 'v', the root in this concordance is adm. One could argue that the root is admh (derived from adm) but I usually list derived roots together with the root they are derived from (though not always). The suffix is tnu, the 't' replacing the 'h' and the nu signifying our. So - the gloss is and our ground.

And if you combine the SimHebrew with the vowels by 'filling in the blanks' you will see how aleph behaves - it's not equivalent to 'a' but only to the glottal stop associated with 'a'. In this case there is one blank too few because the final 'u' of the SimHebrew is a vav in the square text that is really working as a vowel.

So here's a passage to practice on:

ksd-vamt npgwu.
xdq vwlom nwqu.
amt marx txmk,
vxdq mwmiim nwqf.
  • The first word is easy - no prefixes and no suffixes - three strong letters none of which play any role in the grammar. The word is the root. 
  • The second word is v/amt. Another thing you may notice is that roots may come in families. amt is close to amn, and we have seen that n can change into t in some words. These two are also close in sense - truth and faithfulness. 
  • The third word is n/pgw\u. Again the three strong letters pgw give it away, so the n must be doing something to the verb (passive voice is often its role) and the final u is third person plural.
The next line is similar: 
  • xdq, no prefixes or suffixes, 
  • v/wlom (what is that 'o' doing there? Fortunately the concordance is organized on a first two letters basis so it is easy to find out if the vav making an o sound is part of the root or not. See this one and scan it - wl is a very popular beginning to a root. Find wlm to see how many senses this can have.)
  • nwq\u - this time the n is part of the root.
Try the remaining two lines and check your results in the concordance. Notice how when there are exactly three strong letters in the word, it is very likely that this is the root - so the last word (nwqf) must again have n as a prefix. In the second to last word (mwmiim), check if it ever appears in the singular - and note also how the ending ia is the Aramaic equivalent. 

You will, I am sure, be able to decode lots of roots now, but here's a hard one: tt.

What root is the root of tt? Only one letter is left of this root in the word which you will find here. (Look for tt using your browser find key.) Memorization is needed here, and to recognize how some letters change into others for pronunciation. In this example you see the weakness of 'n'. It just drops out of the word. In addition to this, in many roots you will find i's and o's switching with each other or appearing and disappearing within a root. 

To go from root to binyan see this video and it looks as if I could do most of this with an algorithm and so add to the concordance.