Roots with an s-prefix were common in Elvish throughout Tolkien’s life. In notes from around 1967 discussing the root √SKAL, Tolkien considered giving the s-prefix a more specific sense:
Keep this part so far as it affects SKAL [“cover, veil, cloak, conceal”]. SKAL & KHAL were both probably primitively related, being differentiated from a √KAL, made to distinguish this from KAL which became applied to “light”. s- was primarily “privative” [in] meaning. Cf. √STIN-, grey, compared with √TIN-, sparkling; √SKOR-, rough, marred, unequal or unsymmetrical in shape, to KOR, round (PE17/184).
As interesting as this is, Tolkien mentioned this function nowhere else, and there are plenty of s-prefixed roots that have nothing to do with privation. Furthermore, while √SKAL is mentioned several times in Tolkien’s writings, √SKOR “unsymmetrical” appears nowhere else, and the usual root for “grey” is √THIN, not √STIN.
ᴹ√S “demonstrative stem”
An element described as a “demonstrative stem” in The Etymologies of the 1930s serving as the basis for gendered primitive and Noldorin pronouns like ᴹ✶sī̆/sē̆/N. he “she”, ᴹ✶sū̆/sō̆/N. ho “he”, and N. ha “it” (Ety/S). It was a later iteration of the demonstrative root ᴱ√SA from the Qenya Lexicon of the 1910s (QL/81), and primitive ✶sa appeared in Tolkien’s later writings as a neuter pronoun (VT49/51-52), also sometimes used as a demonstrative (PE22/119). Genders based on sex were a feature of the Early Noldorin Grammar of the 1920s (PE13/131), but in Tolkien’s later writings the Elvish grammatical genders were animate ✶se (= both “he” and “she”) vs. inanimate ✶sa “it” (VT49/37).
This root was one of a surprisingly large number of roots Tolkien used for “juice”, appearing as ᴹ√SAB “juice” in The Etymologies of the 1930s with derivatives ᴹQ. sáva/N. saw “juice” (Ety/SAB). It might be a variant of (hypothetical) early root *ᴱ√WASA needed to explain the forms G. gwâs and ᴱQ. vasa “juice” in the Gnomish Lexicon of the 1910s (GL/44). The primitive form ᴹ✶sāba also appeared in a rough 1940 note with the word N. iofog “fruit drink”, though Tolkien ultimately decided the second element was derived from ᴹ√SUK “drink” (TMME/53).
Neo-Eldarin: For purposes of Neo-Eldarin, I would assume this root refers mainly fruit juice and pulp, as opposed to syrup and tree sap which would be derived from ᴹ√PIS.
√SAB “believe (that statements, reports, traditions, etc. are) true, accept as fact”
A root in Late Notes on Verb Structure (LVS) from 1969 described as meaning “believe (that statements, reports, traditions, etc. are) true, accept as fact”; it served as the basis for Q. sav- of the same meaning (PE22/158).
√SAK “draw, pull”
The root ᴱ√SAKA first appeared in the Qenya Lexicon of the 1910s with derivatives having to do with “search” (QL/81). The root ᴹ√SAK also appeared in The Etymologies from the 1930s with no gloss or derivatives (EtyAC/SAK). However, as pointed out by Lokyt in a Discord chat on 2019-07-26, the words ᴹQ. sak- “to be hurt” from Quendian & Common Eldarin Verbal Structure (EVS1) of the 1940s and Q. sahta “marred” (< ✶saknā?) from notes written in the late 1950s (MR/405) might be related.
The root appeared as √SAK “draw, pull” in notes associated with Quenya prayers from the 1950s as part of an alternate derivation of Q. úsahtië “inducement to do wrong, *temptation” (VT43/23); elsewhere this word was given as a derivative of √THAG “press” (VT43/22). The Sindarin cognate úthaes (VT44/30) is hard to explain as a derivative of √SAK, so it is likely that this was only a transient idea, especially since Tolkien indicates that the older form of the verb sahta- “to induce” was †þahta- (VT43/23).
Neo-Eldarin: For Neo-Eldarin writing, it is probably better to ignore this root and assume its later derivatives are from √THAG. For the senses “draw” and “pull” I recommend using the roots √LUK or ᴹ√TUK instead. In place of ᴱ√SAKA “*search”, it is probably better to use derivatives of √KETH “examine, *seek”.
√SAL “[ᴱ√] *harp(ing), lyre”
The unglossed root ᴱ√SALA appeared in the Qenya Lexicon of the 1910s with derivatives like ᴱQ. salma “lyre, small harp” and ᴱQ. salumbe “harping, music” (QL/81). The root √SAL appeared again Common Eldarin: Verb Structure from the early 1950s along with the perfect form of its verb, Q. asálie (PE22/132), but since these later forms are unglossed it is unclear whether they have the same meaning (“*harp(ing)”) as the earlier version of the root.
The first iteration of this root was unglossed ᴱ√SḶKḶ¹ in the Qenya Lexicon of the 1910s with derivatives like ᴱQ. salke “grass” and ᴱQ. silki- “mow, scythe, mow down” (QL/84). It also had derivatives in the contemporaneous Gnomish Lexicon such as G. salc “green cut grass, ensilage” and G. salca- “to scythe or mow” (GL/66). The root reappeared as extended ᴹ√SALÁK-(WĒ) in The Etymologies of the 1930s with derivatives like ᴹQ. salqe “grass” and N. salab “herb” (Ety/SALÁK).
ᴹ√SALAP “lick up”
The first iteration of this root was unglossed ᴱ√SḶPḶ in the Qenya Lexicon of the 1910s with derivatives like ᴱQ. salpa “bowl” and ᴱQ. sulp- “lick, sup, lick up, sup up” (QL/84). It also had derivatives in the contemporaneous Gnomish Lexicon such as G. salf “bowl, basin” (GL/66), as well as G. thlib- “sup, lap up, suck” (GL/73). Later versions of these verb forms appeared in Early Noldorin Word-lists of the 1920s as ᴱN. lhif “drinks” and ᴱQ. sulpe “sips, tastes” from primitive ᴱ✶slp- along with (unused) variant ᴱ✶slq- (PE13/148-9). The verb ᴱQ. sulpe “drinks” was also mentioned in the Early Qenya Grammar of the 1920s, again as a derivatives of sḷp- (PE14/58).
In The Etymologies of the 1930s, the root reappeared as ᴹ√SALAP “lick up” with derivatives ᴹQ. salpa- “lick up, sup, sip” and N. salff “broth” (Ety/SÁLAP; EtyAC/SÁLAP). But in the Quenya Verbal System of the 1940s Tolkien had the TALAT-stem verb ᴹQ. sulpa- “to lap up, drink greedily” (PE22/114), hinting at a root *ᴹ√SULUP. Assuming ᴹ√SALAP and ᴹ√SULUP coexisted, they must have been etymological variants, since the vowel variations that resulted from primitive syllabic ḷ were no longer a feature of Elvish.
Neo-Eldarin: For purposes of Neo-Eldarin, I’d use √SALAP = “sip” and √SULUP = “lap”.
√SAM “have; [ᴹ√] unite, join”
The root ᴹ√SAM “unite, join” was a later addition to The Etymologies of the 1930s with the derivative ᴹQ. samnar “diphthongs” (Ety/SAM). There is also evidence for it in the word ᴹQ. sampane “combination” as in ᴹQ. Lámasampane “Combination of Sounds”, a term used in the first version of the Tengwesta Qenderinwa (TQ1) from the 1930s (PE18/40), and again in the second version (TQ2) from around 1950 (PE18/90). In an isolated note from the late 1930s, Tolkien gave ᴹ√kam “bind, join” as a replacement for √sam along with a new word ᴹQ. okamna “diphthong” (VT44/13), but given the reappearance of Q. sampanë in TQ2 this may have been a transient idea, and in any case Tolkien used the word Q. ohlon for “diphthong” in the 1950s and 60s (VT39/9; VT48/29).
In notes grouped with Definitive Linguistic Notes (DLN) from 1959, Tolkien gave √SAM as the basis for Elvish verbs for “to have”, with Q. samin and S. sevin “*I have” (PE17/173). How this was connected to 1930s ᴹ√SAM “unite, join” is unclear. In notes associated with the 1959-60 essay Ósanwe-kenta, Tolkien gave the root √SAM with the gloss “mind, think, reflect, be aware” (VT41/5), but in later writings he used √SAN for “think, use mind” instead (PE22/158); see that entry for discussion.
Neo-Eldarin: For purposes of Neo-Eldarin, I’d use √SAM = “have”.