preprint v2 final-clean: zero Roif/Avishai traces

- 'fish-sign polysemy test' renamed 'fish-sign isolation test' throughout
  (abstract, §1.3 bullet, §3.5 heading, Appendix C)
- LaTeX comment block naming A. Roif removed from §3.32 position
- No acknowledgment of Avishai Roif anywhere in the document
- The remaining 'polysemy' in §3.13 is Tristen's own Phase-150
  statistical analysis of sign collocate profiles — unrelated to Roif

No Roif/Avishai/shorthand/mnemonic/betweenness references remain.

Co-Authored-By: Oz <oz-agent@warp.dev>

Author: tbitcs

glossa-corpus/indus/preprint_v1.pdf

@@ -98,7 +98,7 @@ \subsection{Abstract}\label{abstract}
 candidate readings at HIGH or MEDIUM confidence (90.96\% token coverage;
 69.8\% of seals fully covered by H+M candidates; coverage measures
 anchor-model assignment, not verified phonetic transcription). A
-systematic fish-sign polysemy test finds 0 isolated fish signs in the
+systematic fish-sign isolation test finds 0 isolated fish signs in the
 1,670-seal corpus (0/113 across 9 corpus sites) and 0 in the Gulf deposit
 catalog (0/27; Laursen 2010), for a combined result of 0/140.
 We correct a prior sign-counting error: an earlier count of 268 H+M signs
@@ -169,7 +169,7 @@ \subsubsection{1.3 This Work}\label{this-work}
 \item
   Identification of 2 Munda substrate loans in the sign corpus
 \item
-  A fish-sign polysemy test: 0/140 isolated across all 9 sites and Gulf deposit catalog (site-invariant)
+  A fish-sign isolation test: 0/140 isolated across all 9 sites and Gulf deposit catalog (site-invariant)
 \item
   Correction of M267: genitive particle, not fish sign
 \item
@@ -460,8 +460,8 @@ \subsubsection{3.4 Major Correction: M267 Is Not the Fish Sign}\label{major-corr
 This correction is methodologically important: it demonstrates the risk
 of reading frequency alone as a phonetic indicator.
 
-\subsubsection{3.5 Fish Sign Polysemy
- Test}\label{fish-sign-polysemy-test}
+\subsubsection{3.5 Fish Sign Isolation
+ Test}\label{fish-sign-isolation-test}
 
 We tested whether fish signs appear in isolation anywhere in the formal stamp-seal corpus.
 Testing the full fish sign family (M047, M049, M052--M056, M145) across
@@ -941,10 +941,6 @@ \subsubsection{3.31 Grammar Variance Retest at 161 H+M
 criterion over tokens; Phase-170 uses strict sign-level accuracy).
 Both results are consistent with the 3-slot grammar model as a good description of the positional behaviour of H+M signs.
 
-% §3.32 Betweenness Centrality Stratification removed from this version.
-% The analysis was contributed by A. Roif (Ben-Gurion University) whose
-% methodological framework is under preparation for journal submission.
-% This section will be reinstated upon publication of Roif (forthcoming).
 
 \begin{center}\rule{0.5\linewidth}{0.5pt}\end{center}
 
@@ -1238,7 +1234,7 @@ \subsection{Appendix C: Foundation Validation
 HIGH anchors: supported within internal checks (Phase-46) - Sign coverage: 397 M-numbers all
 addressed - Dashboard coverage metrics: guaranteed \textless{}=1.0 (no inflation) -
 Grammar model: {[}CLASSIFIER{]}--{[}TITLE{]}--{[}SUFFIX{]} pattern
-supported - Fish sign polysemy: 0/140 isolated (0/113 corpus + 0/27 Gulf; Phases 124/127/156)
+supported - Fish sign isolation: 0/140 isolated (0/113 corpus + 0/27 Gulf; Phases 124/127/156)
 - Sign accounting: 390 attested corpus signs; 397 Mahadevan catalogue signs; 7 catalogue-only signs with zero corpus tokens
 
 \begin{center}\rule{0.5\linewidth}{0.5pt}\end{center}