Electrophilic aromatic substitution: directing groups vs the ranking-edge spec — chemistry scout,
What’s claimed and how well supported
- Labels on this page
- SPECULATIVE ×3
- Strongest label used
- SPECULATIVE
- Main caveat (quoted)
- Setting it beside Navajo and the phosphorus case is a ◆ SPECULATIVE test against a specification, not a correlate, and not a joint field.
Counted automatically from the labels printed on this page. Counts are a reading aid, not a grade of the page.
Date: 2026-10-06 PT (weekday research push, evening)
Status: Chemistry scout. The chemistry is same-domain description from the textbook and primary sources opened below (REAL inside physical-organic chemistry). Setting it beside Navajo and the phosphorus case is a SPECULATIVE test against a specification, not a correlate, and not a joint field.
Audience: Grant (linguistics; less chemistry).
Does not: claim that a directing group, a ring position, or the electrophile is a theta role, an agent, a patient, an inverse or a fourth person; supply a stand-in for a theta role; invent percentages, rate factors, or structures; claim a Zipf result; revise Bridge A.
Why this note. Open question 3 of Zipf fillers in chemistry flagged the benzene directing rule (its §4.2) as "the closest thing in this run to 'the thing that picks a slot'": a rule stated over a class of group (electron donor vs withdrawer), not over atom identity, applying inside an event, with a stated ranking when two directors compete. That note had only the Wikipedia summary. This note opens textbook and primary sources and runs the rule point by point against the six-point ranking-edge spec, the same way TBP apicophilicity ranking edge ran five-coordinate phosphorus.
The case, in plain words. Benzene is a ring of six carbons, each carrying a hydrogen. In electrophilic aromatic substitution (EAS) a positively charged or electron-hungry group (the electrophile, e.g. NO₂⁺ in nitration) comes in and a hydrogen leaves as H⁺. If the ring already carries one group, the new group can go next to it (ortho, two such positions), one further away (meta, two positions) or straight across (para, one position). The group already there decides, mostly, which: some send the newcomer ortho/para, others meta. It also makes the whole reaction faster (activating) or slower (deactivating). With two groups already present, they can agree or disagree about where the next one goes.
What is new (one sentence). The directing rule fits the event half of Grant's shape and has a real second scale — in one event (an entering group replaces a leaving H⁺) groups already on the ring rank the free positions by a class property (lone pair or δ⁺ atom next to the ring), a separate activation scale (–OH ~1000× faster than benzene, –NO₂ >10⁷× slower) decides between two directors that disagree, bulk degrades the ortho share step by step until it reaches zero (tert-butylbenzene + tert-butyl chloride: "no ortho alkylation"), and an occupied position can be attacked (ipso) with an outcome that depends on whether its occupant can leave — but what is ranked is where on the ring one role lands, not which participant takes which role; the rankers are bystanders, not the event's participants; the "impossible" arrangements are near-zeros or route-relative (between two meta groups "rarely", "must … be prepared by some other route"); and no substituent class sits outside the scales, so point 6 leans fail.
- 1. Sources opened this run
- 2. Sourced chemistry
- 2.1 The event and its participants
- 2.2 Scale A — direction class (which positions)
- 2.3 Scale B — activation (how fast), separate from direction
- 2.4 Scale C — bulk
- 2.5 Two directors on one ring
- 2.6 Regime change: kinetic vs equilibrium control
- 2.7 Ipso attack — the electrophile hits an occupied position
- 2.8 Bans at the level of the whole event
- 2.9 Switching the ranking off
- 3. Against the spec (points 1–6)
- 4. Separation
- 5. Verdict
- 6. Open questions
1. Sources opened this run
| Source | How read | Used for |
|---|---|---|
| McMurry, Organic Chemistry (OpenStax, 2023), §16.4 "Substituent Effects in Electrophilic Substitutions" | Web text (HTML saved; figures not seen). Page carries an OpenStax notice restricting ingestion into LLM / AI offerings — quotes kept short, attributed | Activation magnitudes (–OH ~1000×; –NO₂ >10⁷× slower); Table 16.1 nitration o/m/p for 14 substituents; three classes, "There are no meta-directing activators"; halogens split; inductive vs resonance; δ⁺ atom → meta |
| OpenStax §16.5 "Trisubstituted Benzenes: Additivity of Effects" | Web text | Rules 1–3: reinforcing directors; opposing → "the more powerful activating group has the dominant influence, but mixtures of products are often formed"; between two meta groups "rarely occurs … too hindered"; three-adjacent rings "must therefore be prepared by some other route" |
| OpenStax §16.3 (Friedel–Crafts) | Web text | Friedel–Crafts "don't succeed" on rings with a strong withdrawer or a protonatable amino group; "No reaction occurs" |
| Roberts & Caserio, Basic Principles of Organic Chemistry 2nd ed. 1977 (LibreTexts §§22.5, 22.6, 22.7; author permission for research reproduction) | Web text; Table 22-5 and all schemes are images, not visible | "No groups are known" that are meta-directing and activating; kinetic control → most stable intermediate; statistical o:p 2:1; steric hindrance; tert-butylbenzene "gives no ortho alkylation with tert-butyl chloride"; opposing directors "not always easy to predict" (–NHCOCH₃ over –OCH₃); equilibrium control gives 1,3,5- over 1,2,4-trialkylbenzene; ipso attack, methyl cannot leave, NO₂ migrates, ipso share undetectable; p-cymene ~10% ipso product, isopropyl leaves as propene |
| Reusch, Virtual Textbook of Organic Chemistry (LibreTexts "Map: Smith" §16.13) | Web text; rate-factor illustration is an image | Statistical 40/40/20; nitration splits toluene 58.5/4.5/37, tert-butylbenzene 16/8/75, chlorobenzene 30/–/70, ethyl benzoate 22/73/5; "the product-determining step … is the first step"; methyl stabilizes more than more substituted alkyls |
| Olah, Lin, Olah & Narang 1978, PNAS 75: 1045–1049 (PMC411402) | Abstract only (full text is a scanned PDF, not read) | Ortho/para regioselectivity of toluene "independent of the reactivity of the nitrating agent"; sterics "important but not the sole reason"; substrate and positional selectivity "determined in two separate steps"; free-radical nitration "nearly statistical" |
| Kromann, Jensen, Kruszyk, Jessing & Jørgensen 2018, Chem. Sci. 9: 660–665 (PMC5869546; RSC page bot-checked) | PMC full text (figures not inspected) | Site ranking computed per position from protonated-intermediate free energies; 90% / 96% agreement over >525 halogenations at 1 / 3 kcal mol⁻¹ cutoffs; two sites within 1 kcal → bis-halogenated products in ~40 cases; one miss attributed possibly to steric hindrance; fails in 13 cases (<3%) |
| Proceedings of the Chemical Society 8 (1892), pp. 39–40, item 112 (Crum Brown & Gibson; discussion) | archive.org OCR text; page image not checked | The original rule is stated over a property of the group's hydride (directly oxidizable to the hydroxide → meta; not → ortho/para), not over atom identity; Armstrong in discussion: the rule does not apply to dimethylaniline, "which yields much meta-derivative" |
| Wikipedia, "Electrophilic aromatic directing groups" (oldid 1366655643; saved midday) | Saved HTML; secondary, used only beside the textbooks | Competition list: most-activating controls; o/p over m; comparable → sterics; "the position between two substituents, each meta to the other, reacts last"; fluorobenzene para partial rate factor often >1 |
Verbatim passages relied on: excerpts-verbatim.md (not published here) (with the saved HTML/text of each page).
Not reached / not opened (do not treat as read): RSC HTML of Kromann et al. (Cloudflare bot-check; PMC copy used instead). Olah et al. 1978 full text (scanned PDF; abstract only). Crum Brown & Gibson 1892, J. Chem. Soc., Trans. 61: 367 (RSC; not tried — the Proceedings abstract was used). Roberts & Caserio Table 22-5 and Reusch's rate-factor figure (images; numbers used only where printed in text). Baas & Wepster 1971–72 Recueil nitration of alkylbenzenes (Wiley; abstract seen only as a search extract, not used). Moodie & Schofield on ipso substitution (ACS; not tried). Ingold-school partial-rate-factor tables and Stock & Brown's selectivity work (not tried). A Dalal Institute textbook preview (ortho/para ratio, ipso) was downloaded and then deleted unused because its PDF forbids copying.
2. Sourced chemistry
2.1 The event and its participants
- One reaction type: an electrophile replaces a ring hydrogen. Roberts & Caserio: "For all practical purposes, electrophilic aromatic substitution is confined to the substitution of a ring hydrogen." The entering group and the leaving group (normally H⁺) have jobs fixed by the reaction.
- Where the choice is made: Reusch, "the product-determining step in the substitution mechanism is the first step, which is also the slow or rate determining step"; Roberts & Caserio, under kinetic control the preferred product is "derived from the most stable of the possible intermediates." Olah et al. (abstract) split the event further: an encounter / π-complex step "determines substrate selectivity" (which molecule), and the arenium-ion step, "separate for the ortho, meta, and para positions," "determines regioselectivity" (which position). Kromann et al. compute the ranking exactly that way, position by position, from the energy of the protonated intermediate at each C–H.
2.2 Scale A — direction class (which positions)
- OpenStax: "the nature of the substituent initially present on the benzene ring determines the position of the second substitution." Groups with a lone pair on the atom bonded to the ring (–OH, –OR, –NH₂, halogens) direct ortho/para "because of the lone pair of electrons on the atom bonded to the aromatic ring"; any group with "a positively polarized atom (δ⁺) directly attached to the ring" directs meta. Alkyl groups direct ortho/para by an inductive/hyperconjugative effect.
- The 1892 statement is already a class rule, and an odd one: whether the group's hydride (HCl, HNO₂, HCOOH …) can be directly oxidized to the hydroxide (Crum Brown & Gibson, Proc. p. 39). Armstrong recast it as "acid radicles" give meta, and noted the exception of dimethylaniline, "which yields much meta-derivative" (p. 40).
- Splits are graded, not all-or-none (OpenStax Table 16.1): e.g. –NO₂ 7/91/2, –CHO 19/72/9, –Cl 35/1/64, –CH₃ 63/3/34, –OH 50/0/50. Reusch's table gives ranges that depend on the reaction (–CH₃: nitration 55–65 % ortho, Friedel–Crafts acylation 10–15 %). Recorded as printed: toluene nitration is 63/3/34 in OpenStax, 58.5/4.5/37 in Reusch, 58/5/37 in Wikipedia (citing Sykes); conditions are not stated alike.
2.3 Scale B — activation (how fast), separate from direction
- OpenStax: –OH makes nitration ~1000× faster than benzene; –NO₂ makes it "more than 10 million times" slower.
- The two scales are correlated but dissociable. OpenStax: "the directing effect of a group correlates with its reactivity. All meta-directing groups are strongly deactivating, and most ortho- and para-directing groups are activating. The halogens are unique in being ortho- and para-directing but weakly deactivating." The reason given: for halogens the inductive withdrawal (which sets rate) outweighs resonance donation, but the resonance donation, "felt only at the ortho and para positions," still sets direction.
- One cell of the 2 × 2 is empty. OpenStax: "There are no meta-directing activators." Roberts & Caserio: "No groups are known that direct the electrophile to the meta position and, at the same time, make the phenyl derivative more reactive relative to benzene."
- Wikipedia (secondary): fluorine's para partial rate factor is "often larger than one" (activating at para) while ortho and meta are deactivated — the class is not even uniform across positions for one group.
2.4 Scale C — bulk
- Statistical baseline: 40/40/20 (Reusch); ortho:para "would be 2:1" if the positions were equal (Roberts & Caserio).
- Bulk of the group already on the ring: nitration of toluene 58.5 % ortho vs tert-butylbenzene 16 % ortho, 75 % para (Reusch). Bulk of the entering group: "tert-butylbenzene gives no ortho alkylation with tert-butyl chloride" (Roberts & Caserio). Olah et al.: sterics "important but not the sole reason" for the o/p variation.
2.5 Two directors on one ring
- OpenStax Rule 1: if they reinforce, one product (p-nitrotoluene). Rule 2: if they oppose, "the more powerful activating group has the dominant influence, but mixtures of products are often formed" (p-methylphenol brominates next to –OH). Rule 3: "Further substitution rarely occurs between the two groups in a meta-disubstituted compound because this site is too hindered. Aromatic rings with three adjacent substituents must therefore be prepared by some other route."
- Wikipedia (secondary) adds: ortho/para directors "control over meta ones"; when "comparably activating, steric hindrance dominates"; the position between two meta groups "reacts last."
- Roberts & Caserio: with opposed effects "it is not always easy to predict"; N-(2-methoxyphenyl)ethanamide nitrates mainly para to –NHCOCH₃, "which indicates that the –NHCOCH₃ exerts a stronger influence than –OCH₃." (No opened source ranks –OCH₃ against –NHCOCH₃ on the activation scale, so whether this case obeys or breaks Rule 2 is not decided here.)
- Kromann et al.: when two sites are within 1 kcal mol⁻¹, both get substituted in some cases (bis-halogenation, ~40 cases).
2.6 Regime change: kinetic vs equilibrium control
Roberts & Caserio §22.6: AlCl₃-catalysed alkylation of benzene gives 1,3,5-trialkylbenzene "in preference to the expected 1,2,4-isomer" because alkylation is reversible and "one of the least selective"; "Equilibrium control, which favors the 1,3,5-product rather than the less stable 1,2,4-product, becomes most evident when the reaction time, the reaction temperature, and aluminum chloride concentration are increased." So the pattern the directing rule disfavours (all-meta from ortho/para directors) wins once the event can run backwards.
2.7 Ipso attack — the electrophile hits an occupied position
Roberts & Caserio §22.7: - Attack at a methyl-bearing carbon "evidently does occur, but it does not lead directly to substitution products because demethylation, unlike deprotonation, does not occur"; the nitro group moves to the neighbouring carbon, which loses H⁺. - That product "is indistinguishable from that expected by direct electrophilic attack at C2, [so] it is not possible to say how much, if any, product is formed by the ipso route." - In p-cymene (methyl + isopropyl) "about 10%" of the product is 4-nitrotoluene: ipso attack at the isopropyl carbon, and "Unlike methyl, the isopropyl group is eliminated rapidly as propene." At 0 °C further ipso adducts at the methyl carbon are trapped by acetate.
2.8 Bans at the level of the whole event
OpenStax §16.3: Friedel–Crafts reactions "don't succeed on aromatic rings that are substituted either by a strongly electron-withdrawing group … or by a basic amino group that can be protonated"; "No reaction occurs."
2.9 Switching the ranking off
Olah et al. (abstract): "Thermal free radical nitration of benzene and toluene with tetranitromethane in sharp contrast gave nearly statistical product distributions." Same ring, same incoming NO₂ group, different mechanism (radical, not electrophilic); the directing ranking disappears.
3. Against the spec (points 1–6)
Points 1–5 from Inverse ranking §6; point 6 from Fourth person vs ranking §6. Navajo column as sourced in those notes; TBP column from TBP apicophilicity ranking edge §3 (not re-opened). Judgment marks are escaped (*) so they do not render as emphasis.
| Spec point | Navajo (as sourced) | TBP phosphorus (prior note) | EAS directing groups (this note) | Verdict |
|---|---|---|---|---|
| 1. One unchanged event; two participants with fixed roles | Same event; agent and patient fixed | Entering and leaving groups, fixed by the reaction | One reaction type: electrophile enters, H⁺ leaves (Roberts & Caserio). Jobs fixed by the reaction; the position is chosen at the first, rate-determining step (Reusch), separately for o/m/p (Olah, abstract) | Pass for the event, with a difference that matters: what gets ranked is which position the entering group takes, not which participant takes which role. The rankers (groups already on the ring) are bystanders, not participants. Reaction roles are chemists' names, not verb entailments |
| 2. A second ranking scale independent of the roles | Animacy (with definiteness, topicality) | Apicophilicity (plural factors) | Two scales beyond the roles: direction class (lone pair / δ⁺ atom → o/p vs m) and activation (–OH ~10³ faster, –NO₂ >10⁷ slower), plus bulk. Activation is dissociable from direction: halogens deactivate but direct o/p; "no meta-directing activators." Neither scale depends on which group is entering or leaving | Pass, weakened: the scales are properties of bystanders and rank positions; direction and activation are correlated except for halogens; plural factors (induction, resonance, bulk); reaction-dependent splits |
| 3. Scales disagree → marked alignment preferred, default degrades | (62b) inverse good, (62a) direct ?* | Pseudorotation restores the role slot; bulky-ring isomers pass | (a) Two opposed directors: the stronger activator wins, "mixtures … often" (OpenStax Rule 2) — the second scale decides, with a graded outcome. (b) Bulk vs the statistical 2:1 ortho:para: ortho falls 58.5 % → 16 % (toluene → tert-butylbenzene) and reaches zero with tert-butyl chloride. (c) Under equilibrium control the disfavoured 1,3,5 pattern beats 1,2,4 | Partial. Clash-plus-graded-degradation is well sourced, but there is no role-based default being overridden: the "default" is either a count (2:1) or the weaker director's preference, and the "marked" outcome is a different isomer, not a marked form of the same event. (c) is a regime change, not a ranking |
| 4. Some alignments banned outright | (56) *, (60b) *, one bi- per clause | Connectivity bans (cages); monodentate: all 20 reachable | Near-categorical: "no ortho alkylation" (tert-butyl + tert-butyl chloride); methyl cannot be displaced at ipso ("demethylation … does not occur"); position between two meta groups "rarely" / "reacts last" — three-adjacent rings "must … be prepared by some other route"; Friedel–Crafts "No reaction" on deactivated rings; empty class cell (no meta-directing activators) | Partial. The bans come from bulk, leaving-group ability, route, or a gap in the class space — not from the ranking scales themselves, which stay graded. The between-meta case is a route-relative ban (impossible by this event, made another way), the nearest thing to "impossible permutations" found so far |
| 5. Both participants stay inside the unit under the marked alignment | Both remain pronominal arguments | Pseudorotation keeps all ligands | Directors stay on the ring under every outcome of the normal reaction. Exception: ipso with a removable occupant (isopropyl leaves as propene, ~10 % in p-cymene) — there an occupant is expelled | Pass, with the ipso exception noted |
| 6. A way to take one participant out of the ranked set, so the default survives | Fourth person: off the animacy hierarchy; direct survives | Groups with undefined apicophilicity break the frame or pin a slot | No opened source names a substituent class outside the scales. Halogens straddle the scales, they are not outside them. Protonation moves a group across classes (dimethylaniline "yields much meta-derivative," 1892; Friedel–Crafts fails on protonatable amino rings). The only setting where the ranking vanishes and the statistical default returns is a different mechanism: free-radical nitration "nearly statistical" (Olah, abstract) | Leaning fail. No participant-level exit. One event-level near-miss (change the mechanism, the default comes back), which is not the Navajo shape: the fourth person keeps the same event |
Asymmetries (analysis of the sourced cases, not a source's claim).
- Who carries the scale. In Navajo the ranking scale (animacy) is a property of the event's own participants. In phosphorus the apicophilicity scale belongs to all five ligands, including entering and leaving groups. In EAS the scales belong to bystanders already on the ring, and they rank positions for the incoming group; the incoming group's own properties enter only through size (tert-butyl chloride) or mechanism (radical vs electrophile).
- What the ranking decides. Navajo: which participant aligns with which grammatical slot. Phosphorus: which group sits in which of two slot types. EAS: where on the frame one role is realized. That makes EAS a slot-picking grammar over one role — closer to the "which positions get filled" table of Nilova & Stuart (filler note §4.1) than to a voice alternation.
- Direction of the fix. Navajo: the second scale wins, roles fixed. Phosphorus: the role requirement wins (pseudorotation puts the leaving group back apical). EAS: the stronger bystander's preference wins on the activation scale; there is no role default to overturn.
- Trace. Navajo direct and inverse get the same translation (Hale, inverse note). Phosphorus leaves a trace (retention vs inversion). EAS usually leaves a trace (different isomers), but ipso-then-migration erases it: Roberts & Caserio say the share made by that path cannot be measured from the product.
Event-structure grain (description only). The opened sources state the slot choice step by step: encounter (substrate selectivity) → σ-complex at a particular position (regioselectivity) → loss of the leaving group, with ipso branching at the second step (leave, migrate, or be trapped). That is the second opened chemistry case, after phosphorus, where a source tracks who goes where through the sub-steps of one event. Offered as a candidate for Grant's 2026-10-05 framing, not as support for it.
4. Separation
SOURCED (do not inflate)
- Classes and splits (OpenStax §16.4 Table 16.1; Reusch §16.13; Roberts & Caserio §22.5): o/p vs m by lone pair / δ⁺ atom; activation magnitudes; halogens o/p but deactivating; no meta-directing activators.
- Competition rules (OpenStax §16.5 Rules 1–3; Roberts & Caserio §22.6; Wikipedia secondary).
- Bulk: toluene vs tert-butylbenzene splits; no ortho alkylation with tert-butyl chloride.
- Equilibrium control: 1,3,5 over 1,2,4 (Roberts & Caserio §22.6).
- Ipso: methyl not displaced, migration, undetectable share; p-cymene ~10 % isopropyl displacement (Roberts & Caserio §22.7).
- Friedel–Crafts bans (OpenStax §16.3).
- Two-step selectivity; free-radical nitration nearly statistical (Olah et al. 1978, abstract only).
- Per-position computed ranking, 90 %/96 %, bis-substitution near ties, <3 % misses (Kromann et al. 2018).
- 1892 rule stated over hydride oxidizability; dimethylaniline exception (Proc. Chem. Soc. OCR).
HIS SPECULATIVE (not findings)
- Grant's argument-ranking frame and the six-point spec as his shape for a cross-domain edge. Tested here only as a checklist.
- Phrase ↔ molecule / word ↔ atom as the Zipf-shared level. The benzene ring is at his molecule level; nothing here is a Zipf result.
- Language and chemistry both tracking a fine-grained event as it unfolds (2026-10-05). EAS is described step by step in the sources; that is description, not support.
Not claimed
- That a directing group is a theta role, an agent, a patient, animacy, an inverse or a fourth person; that a ring position is a theta slot; that the electrophile is an argument. No source says so; §3 shows how they differ.
- Any percentage, rate factor or barrier not printed in a source opened this run.
- That the –NHCOCH₃ / –OCH₃ case breaks Rule 2 (no opened activation ranking between the two).
- Bridge A is not revised.
5. Verdict
Against the spec: point 1 passes for the event (but the alignment ranked is event-to-position, not participant-to-role); point 2 passes, weakened; points 3 and 4 partial; point 5 passes (ipso exception); point 6 leans fail (no participant-level exit; one event-level near-miss).
Standing among candidates: comparable to phosphorus on 1, 2 and 5; on 3 it is different again (the second scale decides between bystanders, with graded outcomes, but there is no role default to override); on 4 it adds the first route-relative ban (between two meta groups: rare by this event, made by another); on 6 it is weaker than both spinel and phosphorus (no sourced category outside the scales at all). Its distinctive contribution is asymmetries 1–2 in §3: the ranking sits on bystanders and picks positions for one role. Candidate, not correlate.
Plain-words finding for Grant. On a benzene ring the groups already sitting there decide where the next group goes, by what kind of group they are rather than which atom they are, and when two of them disagree the stronger one wins with leftovers going the other way — so chemistry does have a slot-picking rule with a ranking scale, graded losses and near-impossible spots, but it ranks places for a single newcomer, set by bystanders, rather than ranking the participants of the event for roles the way Navajo animacy does.
6. Open questions
- Get numbers for the graded scale: Roberts & Caserio Table 22-5 (image, not visible), Ingold-school partial rate factors, or Stock & Brown's selectivity relationship — whether the activation scale is a single ordering or several by reaction.
- Ipso as a marked path: open Moodie & Schofield's ipso review (ACS; not tried) for how often ipso attack happens where the product cannot show it — the "erased trace" asymmetry in §3.
- The 1892 rule and Armstrong's dimethylaniline exception: open the full Transactions paper (RSC) and a modern account of protonation switching a director's class — is this a participant crossing scales, or leaving one (point 6)?
- Two-director cases with quantified ratios, including the –NHCOCH₃ vs –OCH₃ case, and a source that ranks the two on the activation scale, to see whether "most activating wins" has real counterexamples (a point-3 test inside one molecule).
- Route-relative bans: is there a sourced list of substitution patterns that EAS cannot make directly (1,2,3-, contra-electronic 1,3,5-) beside Nilova & Stuart's frequency table — would that give a "which arrangements are impossible by this event" inventory?
- Grant's multiple-wh / casino test remains with his texts; nothing here bears on it.
Corpus note: eas-directing-ranking-edge-v0.md. The chemistry is same-domain. The comparison to Navajo is speculative and is not a correlate. Not Bridge A. Not a joint field.