# Full IVTFF/Takahashi EVA Structural Experiment

Source: uploaded `LSI_ivtff_0d(1).txt`. Parsed only `;H` lines, identified in the file metadata as Takeshi Takahashi’s full transcription. Page metadata uses `$I` for section/illustration type, `$L` for Currier language, and `$H` for Currier hand.

## Extraction summary

- H lines parsed: **5,207**
- Pages with H text: **225**
- Word tokens after cleanup: **37,967**
- Vocabulary size: **8,071**

## Aggregate features

| group_type | group | n_tokens | vocab | entropy_bits | circuit_rank_mu | successor_entropy_bits | retention | daiin_count |
| --- | --- | --- | --- | --- | --- | --- | --- | --- |
| full | all | 37967 | 8071 | 10.452 | 22675 | 4.361 | 0.008 | 864 |
| currier_language | ? | 3293 | 1699 | 9.832 | 1421 | 1.775 | 0.005 | 37 |
| currier_language | A | 11450 | 3410 | 9.865 | 6561 | 3.382 | 0.009 | 512 |
| currier_language | B | 23224 | 4926 | 9.886 | 13689 | 4.199 | 0.008 | 315 |
| hand | 1 | 7273 | 2213 | 9.333 | 4062 | 3.245 | 0.011 | 384 |
| hand | 2 | 9704 | 2283 | 9.086 | 5493 | 3.762 | 0.010 | 148 |
| hand | ? | 14248 | 4480 | 10.389 | 8292 | 3.240 | 0.007 | 232 |
| section | biological | 6918 | 1550 | 8.579 | 3825 | 3.734 | 0.011 | 84 |
| section | cosmological | 2550 | 1156 | 9.236 | 1218 | 1.979 | 0.007 | 33 |
| section | herbal | 11418 | 3358 | 9.880 | 6635 | 3.372 | 0.009 | 474 |
| section | pharmaceutical | 2579 | 1139 | 9.080 | 1278 | 2.172 | 0.007 | 100 |
| section | stars | 10694 | 3100 | 9.848 | 6307 | 3.330 | 0.008 | 122 |
| section | text | 1626 | 920 | 9.261 | 667 | 1.382 | 0.006 | 28 |
| section | zodiac | 1332 | 808 | 9.082 | 486 | 1.265 | 0.004 | 11 |


## BEH / entropy-direction tests

- **page_Currier_A_gt_B**: nA/1=114, nB/2=82, mean A/1=6.016, mean B/2=6.958, Δ=-0.943, one-sided p(A/1 > B/2)=1.0000
- **page_Hand_1_gt_2**: nA/1=86, nB/2=45, mean A/1=5.840, mean B/2=6.590, Δ=-0.750, one-sided p(A/1 > B/2)=1.0000
- **paragraph_unit_Currier_A_gt_B**: nA/1=141, nB/2=109, mean A/1=5.712, mean B/2=6.453, Δ=-0.741, one-sided p(A/1 > B/2)=1.0000
- **paragraph_unit_Hand_1_gt_2**: nA/1=92, nB/2=63, mean A/1=5.726, mean B/2=6.115, Δ=-0.389, one-sided p(A/1 > B/2)=1.0000

## Matched controls

- **frequency_shuffle** mean over 20 trials: entropy=10.452, μ=24506.2, successor entropy=4.526, retention=0.311%
- **uniform_vocab** mean over 20 trials: entropy=12.818, μ=29941.2, successor entropy=2.394, retention=0.013%

## Daiin diagnostics

- daiin count: **864**
- daiin burstiness: **0.214**
- k=2, period≈18983.5 tokens, power=90075.2
- k=1, period≈37967.0 tokens, power=43195.3
- k=15, period≈2531.1 tokens, power=10496.8
- k=10, period≈3796.7 tokens, power=9567.4
- k=10751, period≈3.5 tokens, power=8752.2

## Top tokens

| token | count |
| --- | --- |
| daiin | 864 |
| ol | 539 |
| chedy | 501 |
| aiin | 470 |
| shedy | 427 |
| chol | 396 |
| or | 367 |
| ar | 353 |
| chey | 344 |
| dar | 319 |
| qokeey | 308 |
| qokeedy | 305 |
| shey | 283 |
| qokain | 279 |
| qokedy | 272 |
| dy | 271 |
| qokaiin | 262 |
| al | 261 |
| s | 255 |
| dal | 253 |