Typo pass: proofread all chapters and notebooks

[DanilSko]

Summary

A full proofreading pass over all 31 Markdown chapters, the README, and the 6 exercise notebooks. Read start-to-finish (in 7 parallel chunks), every candidate re-verified against the actual files before editing, one commit per changed file.

Clear typos (fixed)

Unambiguous spelling errors, grammar/case-agreement mistakes, duplicated words, and copy-paste debris (stray tabs/double spaces, a missing </a> tag, inconsistent trailing punctuation on a repeated call-to-action block). Also two technical/build-breaking slips caught while proofreading the surrounding prose:

• reflection_reflection.md: a [Kapitel 5] link pointed at a non-existent path (corpus_analysis/corpus-analysis_intro) — fixed to the correct docname, confirmed the Sphinx warning disappears.
• corpus-processing_intro-to-nlp.md: {citep}`tiger` used the wrong MyST citation role (the only place in the repo not using {cite:p}/{cite:t}) — fixed to {cite:p}.

Consistency / judgment calls (fixed, called out separately)

Fixed because each is confirmable against the rest of the book or the actual code, not a stylistic judgment call:

• ELTEC-DEU → ELTeC-DEU (5 instances, matching the dominant spelling everywhere else)
• an Hand von → anhand
• relativen Daten → relationalen Daten (a "relative" vs. "relational" mix-up in corpus-processing_nlp-annotation.ipynb, duplicated verbatim in an assessment-quiz feedback string — fixed both)
• Verwaltungswissenschaften → Verwaltungswissenschaft (singular, for parallelism with the other three disciplines listed alongside it)
• A stale folder-name reference (csv → annotations) where the prose no longer matched what the code actually creates
• A handful of similar small drifts between prose and the code/UI it describes (a mismatched notebook-cell name reference, a documented column name that didn't match the actual DataFrame column, etc.)

Noted, not touched

• Book-wide N-Gram/N-Grams/N-Gramm spelling inconsistency, and an Adjektiv-Nomen-Paare vs. Adjektiv-Substantiv-Paare terminology split — both are pervasive and internally consistent within the chapters that use them, so fixing them would need a dedicated pass across many more instances than this one touches.
• Two numeric claims in corpus-analysis_analysis.md that look arithmetically off (a "jedes zwanzigste Wort" example, and a "Prozentpunkte" label that mixes raw-fraction and percentage-point units) — content/pedagogy judgment calls, left for a domain-expert check rather than silently changed.
• Three pre-existing issues, confirmed present on main before this branch (not introduced or touched by this pass): a broken N-Grams link in research-question_operationalization.md, a broken #dateien-schreiben cross-notebook anchor in corpus-analysis_syntactic-ngram.ipynb, and a CRITICAL build error in reflection_reflection.md (an include directive referencing an HTML marker that no longer exists in introduction_learning-outcomes.md). Worth their own issue/PR.

Verified

• verify_content.py (JSON validity for all .ipynb, Python AST parse for all embedded code cells): clean, before and after.
• Full local Jupyter Book fast build (execute_notebooks: off): succeeds, no new warnings introduced by this branch (checked by diffing against a build of unmodified main).